...making Linux just a little more fun!

January 2006 (#122):

The Front Page

By Heather Stern

Herds of wildelaptops roaming free across the African plains


Ubuntu Linux is a Debian derivative distribution of Linux aimed at being fairly easy for someone new to the Linux environment - the ordinary user - and having a somewhat different and thereby hopefully faster version release process. Their motto is "Linux for Human Beings", and Canonical Ltd. in the UK supports them professionally.

This image was inspired by the IRC chat clients found in the Ubuntu default installation (or perhaps their live CD) whose quitline proclaims "set your system free with Ubuntu!" or something like that. It led an IRC friend to joke about herds of wildelaptops wandering free across the African plains...


Did you know that the word "wildebeest" refers to a gnu? Or that the GNU project's sourceforge equivalent is named Savannah? I found the opportunity to wish everyone a visual "Happy GNU Year" too good to pass up.

The African herd that this image describes consists of an image from Dr. Steve Barrett's safari photos (I hope he doesn't mind) and some pictures of various laptops.

Specifically, this shows an Asus 3300, a Toshiba Tecra 8200, and clipart of a Powerbook and something more generic. That's right, kids; Ubuntu has a PPC edition that runs quite nicely on Powerbooks. I've found that the more generic a laptop, the better Linux's chances seem to be... but don't hold your breath about any softmodems. The Blue Wildebeest is a common breed, but the IBM imagery I found didn't seem to suit the angles I needed.

May your gnu year be just a little more fun than last year...

Heather is Linux Gazette's Technical Editor and The Answer Gang's Editor Gal.

[BIO] Heather got started in computing before she quite got started learning English. By 8 she was a happy programmer, by 15 the system administrator for the home... Dad had finally broken down and gotten one of those personal computers, only to find it needed regular care and feeding like any other pet. Except it wasn't a Pet: it was one of those brands we find most everywhere today...

Heather is a hardware agnostic, but has spent more hours as a tech in Windows related tech support than most people have spent with their computers. (Got the pin, got the Jacket, got about a zillion T-shirts.) When she discovered Linux in 1993, it wasn't long before the home systems ran Linux regardless of what was in use at work.

By 1995 she was training others in using Linux - and in charge of all the "strange systems" at a (then) 90 million dollar company. Moving onwards, it's safe to say, Linux has been an excellent companion and breadwinner... She took over the HTML editing for "The Answer Guy" in issue 28, and has been slowly improving the preprocessing scripts she uses ever since.

Here's an autobiographical filksong she wrote called The Programmer's Daughter.

Copyright © 2006, Heather Stern. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 122 of Linux Gazette, January 2006

The Mailbag

HELP WANTED : Article Ideas
Submit comments about articles, or articles themselves (after reading our guidelines) to The Editors of Linux Gazette, and technical answers and tips about Linux to The Answer Gang.

AOL emulator

Sat, 31 Dec 2005 00:59:27 -0500 (EST)
MXRADAR (MXRADAR from aol.com)
Question by tag (tag from lists.linuxgazette.net)
Answered By Mike Orr, Jimmy O'Regan, Pedro Fraile Pé, Jay R. Ashworth

I recently set up a machine with Suse 9.2 and would like to go online with it. I purchased a Creative Labs Model DE-5621 Serial Modem to use with it but am in need of some kind of AOL emulator to be able to tie in to the AOL servers. In experimenting with the machine, I can hear it dial out and attempt connection but that's about it. Any suggestions/souce for this emulation software. Thanks.

[Sluggo] I haven't heard of any AOL client software for Linux. If it doesn't work with the standard 'ppp' package, you may have to try another dialup ISP and access AOL through its website. If you're having trouble with your ppp configuration, post what configuration you have and what exactly you hear from the modem. You should hear a steady tone from both sides, then it will go up and down in pitch, then turn into a hissing sound. But the speaker may be configured to shut off at an earlier point.
I think ppp also has a debug mode that will show all the communication with the server. That may provide a clue where it's blocking.
[Jimmy] I don't think AOL works with straight ppp. Heather mentioned a package that did what the querent is looking for a while back, but the closest I can find is this comment from Ben: "I have not yet run into an ISP to which I couldn't connect (AOL is neither an ISP nor anything that I want to connect to; I'd hate to wash my modem out with soap...)"
The best I can remember is that the site was mostly in French, and the name (shockingly) had something to do with penguins.
[Pedro] Looking into the Ubuntu repositories for something related to AOL, I have found "penggy". The package page shows
connects to AOL via modem or TCP/IP

The project aims to be a full AOL client including all known access
methods like DSL, cable, TCP/IP or modem (with the modem-emulation of
linux via /dev/ttyI* ISDN, too).

Currently, only TCP/IP and modem is supported.
Please bear in mind I have never tried it, nor know really anything about AOL ...
[Jimmy] That's it! The project's website is here: http://savannah.nongnu.org/projects/pengfork
[Jay] AOL connectivity uses PPPoE. If you're stuck behind one with a Linux box, get a compatible router and use that to provide (NATted) IP over Ethernet.
If you want an AOL client for Linux...
well, a version of AOL 5 ported to Linux leaked out the door, and I have a copy once, but I can't find it now, and I can't remember what it was called.
(!) [Jimmy] Does anyone out there have experience with penggy, or the version of AOL 5 Jay mentioned?

wav files from MR8 recorder

Tue, 15 Nov 2005 17:21:29 -0700
Bob van der Poel (bvdp from uniserve.com)
Question by The Answer Gang (tag from lists.linuxgazette.net)
Answered By Jimmy O'Regan

Hi all. A bit of a "forgot my thinking hat" problem here....

I recently purchased a Fostex MR8 digital recorder. I'd not try to go into the recording business with this toy, but for playing around a home with multitrack recording it's a fair bit of fun.

This unit records onto compactflash cards, which I can load onto my linux box from a card reader. No problems here.

And, a program like audacity can read, modify and mix them. Again no problem.

Unless the track is just a partial track with an offset. For example (see the adl file below), I have recorded some tracks from the start of the song. But, 2 tracks are just a little segment in the middle of the song.

See attached rainbow.adl.txt

I know that wav files can have offsets in them (I think), but it seems from looking the the wav file headers that this isn't so here. And, audacity certainly doesn't find the offset.

So, when I load all 5 tracks into audacity tracks 1 and 2 are aligned to the start of the song. They should really start about 40 seconds in. I can move the tracks in the program, but I'll never get them aligned.

[Jimmy] About 42 seconds in? Look at the end of the entries:        _
My guess is the second time field is the length of the track.

Yes, I think that is correct. I have no idea what the stuff after the "L" is.

[Jimmy] Pan settings, maybe, or volume?

Thanks for offering to look up the wav stuff. I'm thinking that if I can get this figured out it'd be easy enuf to write a little program to "fix" the files as they are copied from the CF card. Right now I'm copying all the files, including some junk files. But, it'd be easy to read the text file for the filenames, etc.

[Jimmy] Ecasound would probably be better than Audacity for this: it's a command-line based multitrack editor: http://eca.cx/ecasound

But, wait, the MR8 seems to keep track of stuff in a file. I think that the data in the <EVENT_LIST> section is what I need to look at. But, I'm not sure just what the different fields mean. Anyone with some experience here?

I do know that the files MR80009.WAV, MR80003.WAV and MR80004.WAV are recording of the complete song and MR80011.WAV and MR800012.WAV are partials.

Guess what I'm asking is help in modifying the 2 files to push them into the correct position.

[Jimmy] I'll get back to you on that. I can't think of it off the top of my head, and it's too late to look now, but I'll look at it tomorrow.
[Jimmy] Doh! Use the time shift tool (the one that looks like this: <-> ;) and drag the tracks into place. IIRC, if you set a mark on the timeline before you import the tracks it'll put them at that time.

Yes, I tried that but was unable to align the 2 to the point of making it sound okay. One track was just a bit late or early. Maybe my mouse is to shakey. Unfortunately there doesn't appear to be anything to "align" to. And, even if I got this working once, I'd have to have to do a 2nd time :)

[Jimmy] There is a zoom tool...

Linux as a Windows screensaver

Thu, 22 Dec 2005 21:16:14 +0000
Jimmy O'Regan (The LG Answer Gang)
Question by The Answer Gang (tag from lists.linuxgazette.net)


It uses qemu to boot Linux on a Windows machine. Spiffy.

(!) [Jimmy] Something I've wanted to see ever since I first saw coLinux is a screensaver activated configuration of a cluster-enabled coLinux.
Does anyone care to try their hand at writing an article about this, or some other unusual way of getting Linux running under Windows?


Thu, 08 Dec 2005 16:56:08 -0700
Bob van der Poel (bvdp from uniserve.com)
Question by The Answer Gang (tag from lists.linuxgazette.net)
Answered By Thomas Adam

Hey, I was going to send a message about how pleased I was with my newest toy that "just worked" ...

Got a MIDIsport UNO midi interface. This plugs into a USB port and drives a MIDI keyboard. Got it and plugged it in and (to someone really expecting to do a bunch of manual editing of config files) I found that the alsa drivers found it, registered, etc.

Cool! Now I have a 2nd midi port.

Then I decided to clean up some cable runs and powered off the computer.

When I rebooted, the new MIDI port was gone. Huh??? Checked the cables...and listed /proc/bus/usb/devices ... yes, all there. But, ALSA doesn't know about it.

So, I plugged the MIDIsport in and out ... and it works again.

So, ALSA finds it on hotplug, but not boot. What's going on?

BTW, to this point I have not edited any config files :)

[Thomas] Order of operations, most likely. It's likely that ALSA is loading after hotplug has finished, which means it might not be aware of the device. You can restructure your init sequence, or simply recall /etc/init.d/hotplug to rerun at the end of init, assuming this really is the reason.

That was my first guess. And I've done a bit of testing, etc. and have discovered that the device is found and registered at boot. The interesting part is that if I use alsamixer right after boot it shows the MIDI-Uno-USB device (alsamixer is just grabbing info from /proc/asound/cards). But, the device is NOT there to use. Trying to use it generates an error; and 'aconnect' doesn't show it.

[Thomas] What error?
bob$ aplaymidi -p72 midifile.mid
Cannot connect to port 72:0 - Invalid argument

Not the most helpful :( But, the port isn't there, so it makes sense.

However, plugging the unit in/out again shows it up.

I'm wondering if a modprobe to something might force a registration. I did try 'modprobe usb-midi' but no results there.

[Thomas] No results where? modprobe, if it is successful won't return any output. It's 'lsmod' that will help you here.

I was hoping that modprobe might reset something?

Okay, doing a lsmod before and after the little guy is working does show up something which might point to the problem:

Before a plugin/out (after boot) we have:

soundcore               7008  2 snd,usb_midi
usbcore               108348  7

Afterwards, there is an additional line:

audio                  44032  0

And the others have changed to:

soundcore               7008  3 audio,snd,usb_midi
usbcore               108348  8

Really, just the difference in all this is that the 'audio' module has been added.

Also, there is a file /proc/asound/cards which does not change between no-work and work:

cat /proc/asound/cards
0 [AudioPCI       ]: ENS1371 - Ensoniq AudioPCI
                     Ensoniq AudioPCI ENS1371 at 0xec00, irq 10
1 [Interface      ]: USB-Audio - USB Uno MIDI Interface
                     M-Audio USB Uno MIDI Interface at
   usb-0000:00:10.1-1, full speed

Finally, the file /proc/asound/card1/midi0 has changed from being an empty file to:

bob$ cat midi0
USB Uno MIDI Interface

Output 0
  Tx bytes     : 197
Input 0
  Rx bytes     : 0

More thoughts?

BTW, on Mandriva 2006 there is NO /etc/init.d/hotplug :) I have no idea when this stuff is being done. I think there might be some clues in /etc/hal/device.d


H/W requirement for KDE 3.5 ?

Tue, 03 Jan 2006 14:58:11 +0530
J.Bakshi (hizibizi from spymac.com)
Question by tag (tag from lists.linuxgazette.net)
Answered By Rick Moen, Thomas Adam

Hi list,

does any one of you have the information about the H/W requirement of recently released KDE 3.5 ??

[Rick] Mu. (http://www.catb.org/~esr/jargon/html/M/mu.html)
It makes no sense speaking of KDE 3.5 having particular hardware requirements, since it is a suite of software running at a level far above the level of hardware drivers. If you mean "What's a reasonable minimum amount of RAM to have, so that a default configuration of KDE on Linux will seem OK?", then the answer is probably around 256MB, which is based on experience with earlier releases, but with the knowledge that it doesn't seem to change a lot between releases.
[Thomas] Not to mention that this was answered in a thread on TAG a few issues ago -- asked by Mr. Bakshi himself. :|
(!) [Jimmy] I believe "please share your experiences" in LG #119 is the thread Thomas is referring to.
If it helps, the KDE website does have a requirements page but this refers to compilation requirements.

response to article on scanning

Thu, 29 Dec 2005 05:10:55 +0100
Edgar Howell (Edgar_Howell from web.de)
Question by editor at Linux Gazette (editor from linuxgazette.net)

My article on scanning drew responses from a couple of readers, one, Emil Gorter, who copied you and whose suggestion I find well-done and certainly worth making available for those wanting to archive scans.

And Juhan Leemet politely pointed out that XSANE does, indeed, have the function that I really wanted for use as a copier. He had a problem with it that he hadn't had time to resolve and I then had the same for which I did find a work-around. It is tantamount to using your brakes instead of taking your foot off the gas, so I don't want to suggest it to anyone.

But the bottom line is that using XSANE to copy directly from scanner to printer can work quite nicely.


Scanning and file-size

Fri, 16 Dec 2005 19:15:45 +0100
Emil Gorter (emil from ripe.net)
Question by lg (lg from howell.de)

Hi Edgar,

I read your article http://linuxgazette.net/121/howell.html and it reminded me of a problem I had faced. For my work I had to figure out a way to scan and store contracts digitally. The file-size issue was quite interesting with ++6000 pages to go :-)

I found that creating PDFs is quite easy, and gives very good results. Nice quality, great reduction in size. Example:

 $ tiff2ps -2 -a Scan0001.tif | ps2pdf -sPAPERSIZE=a4 - > Scan0001.pdf
 $ ls -1 --size -h Scan0001.*
 1.1M Scan0001.pdf
  25M Scan0001.tif

By the way, for practical reasons the scans were made with a Windows machine.. Still have to try SANE myself someday.

I hope this helps!

Emil Gorter

"Summer of Code" projects list updates

Tue, 08 Nov 2005 06:39:48 +0000
Jimmy O'Regan (The LG Answer Gang)
Question by ()

First of all, I would like to offer my most sincere apologies for the negative way I portrayed the WinLibre installer projects. What I took to be a duplication of effort was in fact three separate parts of a larger system. Thanks to Noemi Tojzan for providing this corrected project list:

     Finished projects:
         - installer (2 students)
         - updater (1 student)
         - control centre (1 student)
         - MacLibre (1 student)
         - 3 -original- games (3 students, 1 per game)
         - Final Touch (or Image Manipulation Tool) (1 student)

     Failed projects:
         - CDRoast GUI (1 student)

There were also inaccuracies in my list of Wine's projects. Thanks to Kai Blin for providing a corrected list:

Thanks to Patrick Walton for a list of LiveJournal projects:

Thanks to Meredith Patterson for pointing out that I had (originally) listed Charun mistakenly instead of QBE (http://pgfoundry.org/projects/qbe), to Ivan Beschastnikh for providing the correct title to the Internet2 project I had listed as "User-space transport": FB-FR-CCCP (http://fb-fr-cccp.sourceforge.net), to Jimmy Cerra for pointing out two Semedia projects: SWAPI (http://swapi.ourproject.org) and SPARQL for various RDF databases (http://sourceforge.net/projects/sparql), and to Anil Ramnanan for pointing out an Apache project: an Eclipse plugin for Apache Forrest.

Thanks to Ross Shannon for pointing out these Gallery projects:

I owe a rather sheepish apology to Oleg Paraschenko, one of the Google-mentored students: I had found XSieve, but lost the details when it came time to write my article. Oleg has more than made up for this:

I later found a complete list (http://development.openoffice.org/summerprojects.html) of Open Office projects:

I would encourage anyone interested in "Summer of Code" to have a look at Google's list of projects (http://code.google.com/soc-results.html)

NetPBM equivalent for vector graphics

Sun, 16 Oct 2005 20:55:25 +0100
Jimmy O'Regan (The LG Answer Gang)
Question by The Answer Gang (tag from lists.linuxgazette.net)


Mentioned in a Newsforge story about Xara open sourcing Xara Extreme:

Also in Newsforge today: Finding voice codecs for free software


found something in the attic...

Wed, 28 Dec 2005 17:43:27 +0100
Karl-Heinz Herrmann (khh from khherrmann.de)
Question by The Answer Gang (tag from lists.linuxgazette.net)

Hi I'm working on my Laptop again after a long while during which it was just sitting in the corner. I found something in the old draft folder -- And I've no idea if that ever got sent. So here you go....

Oh and a happy new year to everybody.


On Tue, 25 May 2004 09:38:38 -0400
Ben Okopnik <ben@callahans.org> wrote:

> I'm writing to you because you're the "quiet bunch" :)

thanks ;-) I was following the "linuxgazette.net down" thread with half an eye.

> Right at the moment, we've got a TAG discussion going on where I've
> asked for input from everybody on the above topic. To restate it,
> I'm looking at ways to bring LG onto a parallel course with Linux
> itself: the OS is booming, but we're lagging. From my own perspective,
> I see a need to modernize a bit; to consider the types of people now
> entering the world of Linux (corporate and business users) and to
> address their needs. "Amateur Linux user" does not, these days, mean
> "amateur computer user".

lets start with history: I started out in TAG when many questions came it on general install questions like partition setup, lilo tricks to make it work, lost root passwords. As TAG is no newsgroups with expiring articles after some time we had perfect in depth solutions to everything in this category of FAQ. The level of discussion in TAG was slowly moving up in complexity and/or detail of a particular problem. I got the feeling even before the com->net switch that there is a little stagnation in worthwhile topics and things I/we can answer. At the beginning there were questions e.g. on network setup on a level I could join in at least answering in part. It's quite a while that I've seen a question I had any chance answering.

Others have suggested finding out wher new and current Linux users and LG readers are and pick them up at their level and target their interests. Pretty good idea, but I rarely use a GUI so I can not speak of first hand experience. I just installed the brand new SuSE 9.1 which came along with a collection of spiffy K* applications from Mind mapper to notes keeping and e.g. GTKtalog to catalog CD contents. Important to me right now as I've a growing CD collection in CD storage bags and I spend more and more time searching for a particular one. Well -- I read in a bunch of CD's, program crashed twice so I pressed the save button frequently. GUI interface is nice, pop in CD press a button, CD content is archived. Next CD. A few days later I tried to open the archive -- program segfaults. I've still the XML export of the data.... rather unreadable without a conversion program/frontend. My next try will certainly be a bunch of "ls -R" files which I can grep. But it seems this kind of solution is nothing we can sale to "the new TAG reader" as they want "aesthetic solutions" as sombeody put it.

I agree with that analysis BTW: from talking to people either having a Linux on some second partition to considering installing one, they want the perfect solution, GUI, everything away only a few mouse clicks. And they want pretty looking neat applications for everything.

My problem with that is: I don't have the answers, as the "nice neat applictions" might not be there or at least I don't know of them. There might be a need of new TAG/article writers for a new type of readers.

So the task at hand might be to attract new members to LG writing for these new readers.

Many topics I was concerned with lately (Wlan, AMD64) are highly specific, involve kernel patches and or/special distributions which are not suitable for new comers to Linux.

Which leads me to a little question anyway: Is LG to be for newbies?

An idea anyway: By talking to people (here in Germany that is) more and more are annoyed about the spam flood. It might hit here slightly delayed compared to US, or maybe not. Well I was searching a bit for Win solutions as all of them run WinXX. No luck with freeware or even open source (portable, cygwin) solutions. So my standard answer is getting "well, since you won't consider Linux pay up for a commercial spam filter".

An article on how to use a Pentium classic PC from the attic and turn it into a Spamfilter-firewall in a recipe (starting out at the point where you insert a knoppix or install CD) might be interesting to some people.

This page edited and maintained by the Editors of Linux Gazette
HTML script maintained by Heather Stern of Starshine Technical Services, http://www.starshine.org/

Published in Issue 122 of Linux Gazette, January 2006

More 2 Cent Tips!

See also: The Answer Gang's Knowledge Base and the LG Search Engine

Just Bad and Wrong... yet cool

Rick Moen (LG Contributing Editor)
Question by ()
Answered By Jimmy O'Regan, Ben Okopnik

"bash is terrifying me again":

Bash pseudo-sockets: Oh my!

followed by: "Dude, you've got a smiley in your prompt."

[Jimmy] Followed by:
:) fishsticks$ false
:( fishsticks$

[Raj] How did he do that. I want something like that too. If anyone knows, do tell.

jimregan@jimmy ~
$ echo $PS1
\[\033]0;\w\007 \033[32m\]\u@\h \[\033[33m\w\033[0m\] $

jimregan@jimmy ~
$ smiley () { if [ $? == 0 ]; then echo ':)';else echo ':(';fi; }

jimregan@jimmy ~
$ PS1="\$(smiley) \h$ "
:) jimmy$ false
:( jimmy$ true
:) jimmy$
Um... I should probably explain that.
$? holds the exit status of the last process: zero for success, non-zero for failure; $PS1 is the primary prompt.
It's \$(smiley) because you want bash to evaluate it every time: $(smiley) would just evaluate it at the time you set the variable, as would putting the contents of the smiley function into PS1
\h is the shorthand for \$(hostname)
[Ben] Darn it, Jimmy, you always beat me. What, you have a fast connection or something? :)
[Jimmy] My connection, in "the cheque is in the mail" speed is 115.2 Kbps, so no.
[Ben] My way was just slightly different:
# More-or-less undocumented Bash-ism - a C-style trinary operator
smiley () { echo -e ":\\$(($??50:51))"; }

BSS and Data Segment

Mike Zheng (mail2mz from gmail.com)
Question by tag (tag from lists.linuxgazette.net)
Answered By Lew Pitcher

Hi All,

For a program, the BSS contains un-initialized variables, Data contains initialized variables. Why do we want to separated these two categories of variables?

[Lew] It's an optimization thing. If each binary contained an image of 'data' that included the uninitialized variables, then the binary would be larger in size than if the image did not include uninitialized variables. If you don't map these variables to your 'load image' (the contents of the binary), but instead, map them to memory at execution time, your binaries can be smaller without a loss of functionality.
The 'typical' (conceptual) memory map (once a program has been loaded into memory) looks something like this...
          |  stack space   |
          |       :        |
          |       v        |
          |                |
          |                |
          |                |
          |       ^        |
          |       :        |
          |   heap space   |
          | .. .. .. .. .. |
          |      BSS       |
          |      DATA      |
          |                |
          |     (CODE)     |
          |                |
(on most implementations, CODE is kept in a separate address space from DATA/BSS/heap/stack)
Heap 'grows' up from the end of BSS towards the highest address Stack 'grows' down from the highest address towards the end of the BSS Heap allocation code keeps a boundary between the top end of the heap and the bottom end of the stack.
The 'loadable' part of the binary is the CODE and DATA part. BSS is how the loader accounts for the room needed to hold the uninitialized data.

Some program clear the BSS by itself, is it necessary?

[Lew] A mere convenience, nothing more.

email issue

J.Bakshi (hizibizi from spymac.com)
Question by tag (tag from lists.linuxgazette.net)
Answered By Jimmy O'Regan, Rick Moen

Hi list,

At first I wish you all a very ******HAPPY NEW YEAR******.

Here is a technical query. Frequently I get some advertisement-emails with EMPTY "form" header & EMPTY "to" header. Even the "detail-header-view" of sylpheed-claws can't show any header information. What is the technology which makes this possible and how can we prevent such emails ?

[Jimmy] The technology that makes this possible is called [drumroll]... SMTP!
To see how it works, use telnet:
$ telnet smtp.o2.ie 25
220 smtp1.o2.ie -- Server ESMTP (Sun ONE Messaging Server)
MAIL FROM: test@example.com
250 2.5.0 Address Ok.
RCPT TO: jimregan@o2.ie
250 2.1.5 jimregan@o2.ie OK.
354 Enter mail, end with a single ".".
Subject: Test

A test
250 2.5.0 Ok.
221 2.3.0 Bye received. Goodbye.
Now, I check my mail. In Mozilla Mail, it shows up as this:
Subject: Test
Date: 11:45

A test
View->Message Source:

See attached mbox.txt

What'll be interesting now will be to see how many mail apps that use mboxes are broken by that } :)
[Rick] J.Bakshi's other question was:
...how can we prevent such emails ?
The answer is [drumroll]... don't use SMTP e-mail.
Or alter your receiving SMTP server's filtering rulesets to reject or drop incoming mails not meeting certain technical requirements, including possession of key header lines. Generally, this requires running your own SMTP host (mail tranfer agent) -- rather like A.J. Liebling's dictum that freedom of the press is limited to those who own one.

Thanks to both Rick Moen and Jimmy O'Regan for their explanation. Jimmy's answer is a bonus as he explained through a good practical example. I have just come to know mixmaster. Though haven't gone through its documentation yet, but aptitude show mixmaster shows that it can prevent the recipient knowing your email address.

[Rick] Er, you might have a use for mixmaster, but the preceding discussion didn't seem to suggest that, and I fear that you might be misunderstanding its intended purpose.
Mixmaster is a client-server software tool for people running MTAs (SMTP machines), and their users, with the client piece allowing the users to issue pseudonymous e-mails, which the mixmaster daemon (server) piece then remails on their behalf.
It would be very strange to send your routine, normal outgoing mail through Mixmaster, specifically because it would greatly obscure the identity of the sender. Further, and more to your original point, it would do nothing to prevent you from also receiving spammer-and-malware generated junkmails at your real e-mail address, such as the garbage mail you cited that had a couple of null interior headers.
As Ben knows, I've long been of the opinion that pretty much all "hide from spammers" strategies (including, well, "preventing the recipient knowing your address") are a dumb idea, really don't work worth a damn, and interfere with legitimate Internet usage to a degree I find unacceptable. At least for those of us who control our own MTAs, there are effective countermeasures, that don't involve self-concealment in any way.

Thanks Rick. You have provided a short as well as good article on Mixmaster & its application ( I may be permitted to say for beginners ). Now I can under stand that as a Workstation owner ( and not an MTA) I don't need Mixmaster. Thanks again.

Running isoLINUX from a REALLY old machine

R.M.Deal (deal from kzoo.edu)
Question by tag (tag from lists.linuxgazette.net)
Answered By Ben Okopnik

I just read the enlightening article Booting Knoppix from a USB Pendrive via Floppy by Ben Okopnik from the #116 issue of the LINUX gazette.

[Ben] Well, thank you for the compliment - and I'm glad you enjoyed the article.

I have a similar problem but with an older machine, a Toshiba Satelite Pro 410 CDT, for which I have an external floppy drive. My problem is that the BIOS does not give me the option of booting from a CD. I can boot from either the hard disk or from the floppy.

[Ben] Have you looked for a BIOS update? I'm not saying that one is definitely available, but if it is, then that would be the simplest solution. Take a look at Toshiba's site and see if there's a download available.

Great suggestion. I shall do that.

Now while I found one article on booting Knoppix from floppies, I wonder if it would be a big deal to modify the script developed by Ben to produce a Knoppix boot, not for a USB pen drive (this portable has NO USB port) but from a floppy. By the way, the Toshiba has installed as an operating system Windows 95 and has only 16 M of RAM so running even KNOPPIX may be a challenge.

[Ben] The problem is that, shortly after I wrote the article, the information in it became outdated: the next release of Knoppix had a minimum size for the kernel/modules/etc. that was much too large to fit on a floppy. However, for your application, there's an easy answer - or at least used to be (I don't have a Knoppix CD handy to check): just fire up your Wind0ws, look at the Knoppix CD, and there should be a batch file there called "mkfloppy.bat". Run it; after it writes a boot floppy for you, simply reboot, leaving both the floppy and the CD in the machine.

Well, almost. I could not get anything under my Windows boot on the portable (a laptop only if you have strong quadriceps) so I booted my LINUX system (suse 9.3) with KNOPPIX and couldn't find a mkfloppy.bat there.

[Ben] That wouldn't have helped anyway, since all the commands in the batch file are DOS programs.

Right, but I could not find the batch file from W95 or suse.

However, in a file knoppix-cheatcodes.txt, the procedure to handle floppy boot only is described, using a LINUX command in KNOPPIX named mkbootfloppy. However, when I boot with KNOPPIX, I can find no such file, only a mkboot. That mkboot command looks right but when I run it with an installed blank floppy, after writing on the floppy drive, I get an error message, in German. Now I am living now in Germany (Weimar) and am taking a course in German but I do not have the resources to fully translate technical German.

[Ben] Whoops... Perhaps you could post that error? My technical German has risen above zero (in part, due to Knoppix :), and there are several German-speaking folks here who could help.

OK. Attached is a script of what happened when I tried to produce the boot floppy. It was interesting making the script in KNOPPIX and transferring it to /mnt/hda2/home/ralph/ after mounting /mnt/hda2. Please note that I am working with Knoppix V3.8.1 (2005/04/0 8). I am not updating right now because I cannot get DSL in Weimar despite having it in Munich and so am having to use ISDN (hence my switch from Fedora to Suse) in which I can get no flat rate! Once DSL is available ("tomorrow"), I'll update my version.

Script wurde gestartet: Mo 21 Nov 2005 10:21:56 CET
root@4[~]# mkboot

Insert a floppy diskette into your boot drive, and press <Return>.

Creating a lilo bootdisk...
cp: Schreiben von ,,/tmp/boot15663/vmlinuz": Auf dem Gerät ist kein Speicherplatz mehr verfügbar
root@4[~]# exit

Script beendet: Mo 21 Nov 2005 10:23:11 CET
[Ben] "No space left on device". At least the version with the umlauts, etc. - "Auf dem Gerät ist kein Speicherplatz mehr verfügbar" - means that. Seems that it's not all that uncommon: searching the Net produces a number of hits. It seems that in some cases, this is produced by bad floppies - but in other cases, it's a result of bad communication with an external floppy drive.
I'd suggest making the floppy from your Wind0ws, just to see if it works. Having two OSes can be handy for resolving "is it hardware or software?" questions.

I plan on working on deciphering that error message but meanwhile I took the floppy that resulted from that action and tried to boot my portable with it. Unfortunately, it only booted into W95; it seems the boot floppy is faulty. You referred to two floppy disks and so did the cheatcodes.txt file but there was no prompt to put in a second floppy. It may be time for me to contact the programmer of KNOPPIX directly to see what the problem is. I'm guessing that the cheatcodes.txt file was written for an earlier version of KNOPPIX and doesn't describe the procedure correctly for the new version. Did I say LINUX was fun??

[Ben] Sure. Part of the fun is the ability to make things happen that aren't the norm. :)

Second, related, question: does anyone recommend a version (old if necessary) of LINUX that would work well as the operating system on the Toshiba?

If these are in some FAQ, I'd appreciate a link.

[Ben] I find that Puppy Linux (http://www.goosee.com/puppy) works well on older machines, although I haven't done a broad range of experimentation with it; it has a nice desktop and a sufficient range of installed applications that I don't find it restrictive. Other folks here may have other suggestions.

Ain't LINUX fun, Ralph

[Ben] [grin] Actually, yeah. I like it, m'self.

Tuppence Tip: URL scraper.

Thomas Adam (thomas from edulinux.homeunix.org)
Question by TAG (tag from lists.linuxgazette.net)

Hello --

This is really an on-going issue from my post regarding urlview and the logging of URLs. I've since decided to take a different approach, and thus far this method works quite nicely. I'm now using 'multi-gnome-terminal' (MGT), 'multitail', 'gmrun', plus a helper script. The overall aim of all of this, was to be able to:

You might wonder what's so hard with this -- the problem is that X11-forwarding on my server takes forever -- it is only a poor P166 with 64MB of RAM, after all. The irssi session resides on the server, so I needed a way of pseudo-opening the URLs as though the request originated on my workstation.

It's unfortunate that I have to use MGT, since it is a memory hog, but needs must. I use it because that has the ability to automatically hotlink URLs -- so that actions can be assigned to it when they're clicked on. Based on this premise, filling in the gaps was easy.

I mount my server's filesystem via 'shfs' -- which I'm now using as a replacement for NFS. I really like it (and a lot less buggy than lufs, and its ilk). This way, I can use multitail to keep an eye on that file. The url logging script that I use from irssi, is "url_log.pl" [1]. All of the logged entries are in the format:

Sat 10 Dec 2005 00:57:08 GMT nick #chan URL

... and I wanted multitail to colourise the output, as it does for other files. That was easy -- just create a new colourscheme for it in /etc/multitail.conf:

cs_re:green:^... .. ... ....
cs_re:magenta:..:..:.. ...
cs_re:red:.* \#fvwm (http|https|ftp)://.*$
cs_re:blue:.* \#bash (http|https|ftp)://.*$
cs_re:cyan:.* \#elinks (http|https|ftp)://.*$
cs_re:yellow:.* \#hug (http|https|ftp)://.*$

So to break this down a bit, remember a typical entry from this file will look like:

Sat 10 Dec 2005 00:57:08 GMT nick #chan http://myfoo.com

Hence: "Sat 10 Dec 2005" will appear in green. "00:57:08 GMT" in magenta, and the rest of the line will appear in whichever colour is matched by the channel the URL was quoted in. So, it looks pretty. :)


... should obviously be changed to match whichever file is going to hold the urls from the url_log,pl script.

The next stage was to determine what happens when I clicked on a URL (I say click -- the shortcut to opening a URL via MGT is 'CTRL + middleclick'). I didn't want everything to be sent to my browswer. This is where the "gmrun" utility comes in useful [2]. For those of you that have never used it, it's a very handy, and customisable tool. One of the features it has, is pre-defining prefixes for certain applications. So for instance, I could enter into gmrun:


... and depending on what I had told gmrun to do with the 'man' prefix, it would open up the bash man page. Neat, eh? So I wanted to have a separate program to open up images, and URLs (it's quite often the case that people post links to screenshots, that I don't want to open in a browser, but would just rather 'see'). I needed to use a helper script to do this, as gmrun accepts no command-line options. The trick I used (in order to make it appear directly in the gmrun window, as though I had typed it), was to append it to gmrun's history file --- if set correctly, gmrun will display the last entered command. No biggie, here it is:

See attached runvia.sh.txt

(Saved as ~/bin/runvia.sh -- and chmod 700 ~/bin/runvia.sh)

So, I'm able to flag to gmrun that if the URL I am clicking on is an image, then tell it so, else, flag it to open up in elinks (this is my primary browser -- although I wanted a specific handler for it.) But in order for that script to process the URL that we clicked on from MGT, we need to tell MGT to perform that action. This is easier than you'd think, and involves editing the file: $HOME/.gnome/Gnome, such that:

default-show=runvia.sh "%s"

Going back to gmrun, we lastly need to tell it what acrtion to take for the 'elinks:' and 'image:' prefixes. That information is stored in /etc/gmrunrc, although I copy this to ~/.gmrunrc, personally, and edit it, so that for the image handler:

URL_image = sh -c 'feh %s'

('feh' has the ability to read images via http).

And for the elinks handler:

URL_elinks = sh -c '${TermExec} elinks -remote "%s" && FvwmCommand "All (*ELinks*) FlipFocus"'

"${TermExec}" is a variable defined further up in the file that looks like this:

Terminal = rxvt
TermExec = ${Terminal} +sb -ls -e

... and that's it. It seems to be working really well.

Since I use FVWM, I wanted to (when I had decided to click on a URL) to focus the webbrowser -- hence the reason why I'm using FvwmCommand. This is optional of course. Although to continue on a similar theme, the style of the "gmrun" dialogue window is set to the following:

Style Gmrun GrabFocus

.. so that when it pops up, I can hit enter, knowing that the Gmrun window will always have the focus, to execute whatever is inside it.

You can see a screenshot[3] of the url-logger in action.

Hope someone finds this useful, or can derive other ideas from it.

[1] http://www.irssi.org/scripts/html/url_log.pl.html

[2] http://www.bazon.net/mishoo/gmrun.epl

[3] http://edulinux.homeunix.org/~n6tadam/fvwm/ss/url-logger.png

Global home movies with cron and Imagemagick

Karsten M. Self (karsten from linuxmafia.com)
Question by tag (tag from lists.linuxgazette.net)

I was just showing this to Heather. I've been trying to figure out what's going on with our strange California weather (tropical rainstorms for the past two weeks) and decided that moving satellite images were more interesting than static ones. Among other discoveries, we've been getting a tropical express mainlined to us over the Pacific for the past several weeks, and the movies make this very apparent.

There are two parts to this tip: first, you want to start grabbing satellite images as they're available, then you want to string them together to view them. Updates are 30 minutes and 3 hours for regional and hemispheric views, respectively.

For the fetch, I use cron to schedule downloads, storing them in a large local area, 'data/weather', with a timestamp added to the filename. Recipie, add to your personal crontab:

See attached crontab.txt

To animate images, say, from the past five days, you can animate 'find' results. To look at west-coast infraread images:

animate -pause 2 -delay 25 $( find . -mtime -5 -name WCIR\*  | sort)

There's some cleanup you may need to do as the download images occasionally get out of order. I've found deleting duplicates seems to fix this pretty readily.

You can also experiment with image enhancements via ImageMagick, such as increasing size and contrast, and reducing noise, in the downloaded images.

This page edited and maintained by the Editors of Linux Gazette
HTML script maintained by Heather Stern of Starshine Technical Services, http://www.starshine.org/

Published in Issue 122 of Linux Gazette, January 2006

The Answer Gang

Linux Gazette 122: The Answer Gang (TWDT)
...making Linux just a little more fun!
(?) The Answer Gang (!)
By Jim Dennis, Jason Creighton, Chris G, Karl-Heinz, and... (meet the Gang) ... the Editors of Linux Gazette... and You!

We have guidelines for asking and answering questions. Linux questions only, please.
We make no guarantees about answers, but you can be anonymous on request.
See also: The Answer Gang's Knowledge Base and the LG Search Engine


¶: Greetings From Heather Stern
(?)DirectPad Joystick
(?)LN -- have I gone brain dead?
(?)System V or BSD?
(?)Yikes. --or--
Missing libraries
...and how to find them

(¶) Greetings from Heather Stern

(?) DirectPad Joystick

From Triyan W. Nugroho

Answered By: Peter Knaggs

Hello Gang,

Here in Indonesia we have very cheap joystick using DirectPad Pro driver (it's actually just a modified Playstation joystick).

(!) [Peter] Does it connect to the parallel port of your Linux machine? If so, you might find the README and README-parport in the following are helpful to get an understanding: ftp://atrey.karlin.mff.cuni.cz/pub/linux/joystick/joystick-1.2.15.tar.gz
If on the other hand, it connects to the game port of your sound card, don't use the parallel port driver, of course :)

(?) Yes, it does connect to parallel port.

(?) I have use it for a long time in Windows, but still have no idea how to use it with Linux.

(!) [Peter] Did you read these pages? http://atrey.karlin.mff.cuni.cz/~vojtech/input http://www.linuxgames.com/joystick

(?) Yes. But the documentation in the Linux kernel says that the kernel has support for DirectPad joysticks, using gamecon module.

(!) [Peter] Yes, "gamecon" is likely the right module to use.

(?) So I think I don't have to compile the package provided in the abovementioned pages..

(!) [Peter] True. Still, they're useful in that they combine the userland tools and the kernel drivers all in one place with some guides and FAQ and other such. It gives a good starting point, and from there you can work on the specific issues of the 2.6 kernel, once you know how to get the device to work correctly in a plain 2.4 kernel.

(?) And I'm still confused, because the package in these pages use different filename.

(!) [Peter] Yes, well filenames change over time, code hardly changes all that much.

(?) For example, Linux kernel uses joydev.c, but in these pages it is joystick.c. And for DirectPad joysticks, Linux kernel uses gamecon.c, while in these pages it is joy-console.c.

(!) [Peter] Those would have been the names during the 2.4 series kernel.

(?) I have been Googling and found that the Linux kernel actually have support on it. And then I found several documentation in the Linux kernel, but I always fail :(

(!) [Peter] What did you try, and how did it fail, specifically?


cd /dev
rm js*
mkdir input
mknod input/js0 c 13 0
mknod input/js1 c 13 1
mknod input/js2 c 13 2
mknod input/js3 c 13 3
ln -s input/js0 js0
ln -s input/js1 js1
ln -s input/js2 js2
ln -s input/js3 js3
(!) [Peter] Well, those entries would be created by udev if you were using a 2.6.12 or later kernel, once the module was loaded correctly.
It seems, though, that the module "gamecon" isn't loading for some reason. Perhaps because it isn't finding the parallel port device? Or because the parallel port is being used by another module (have you checked that "lp" module isn't loaded)?
Are you sure the parallel port is enabled in the BIOS? Do you happen to have a parallel port printer and test that printing works OK, or some other way to know if the parallel port is actually enabled? You mentioned the joystick worked in Microsoft but it wasn't clear if it was on the same machine.

(?) But every reboot, my js* is always gone  :( How can I fix it?

(!) [Peter] Not sure about that, could be many reasons. It's no harm to create the devices manually, but unless the module is loaded successfully they will just give ENODEV when you try to use them.


root@devel:/home/i2c# modprobe joydev
root@devel:/home/i2c# modprobe gamecon
FATAL: Error inserting gamecon
(/lib/modules/2.6.8-2-386/kernel/drivers/input/joystick/gamecon.ko): No such
(!) [Peter] Have a look in "dmesg" to see if there's any other messages. Also, can you see anything mentioning "parport" in your boot messages, if so can you send it, e.g.:
parport: PnPBIOS parport detected.
parport0: PC-style at 0x378 (0x778), irq 7, using FIFO
That would tell you the IRQ line to look for in /proc/interrupts corresponding to the parallel port.

(?) I've tested on 2.4 and 2.6 series, but no difference...

(!) [Peter] Hmm, so it looks like more debugging will be needed. Have a look in the sources, you can add in some printk(KERN_INFO along the path from module_init(gc_init); so add in some tracing of gc_init and see why the gc_probe function isn't happy. You should see your debug additions come out in "dmesg" when you modprobe gamecon and you can keep adding more debug tracing and doing rmmod gamecon and modprobe gamecon to try again. Be careful though with your code, as you're likely to crash your kernel sometimes doing this sort of thing.

(?) Can you explain step-by-step how to configure the kernel to work with this joystick? I use Debian 3.1 and kernel 2.6.18.

(!) [Peter] I wouldn't have the same hardware as you, so I couldn't provide a step-by-step guide. My guess woiuld be that the joystick stuff might be easier to get to work using a 2.4 series kernel rather than 2.6 series. I haven't been following the development in the 2.6 series much on joysticks.

(?) My sound card (Intel i810 onboard) doesn't work on 2.4, so I have to use the 2.6 series :(

(!) [Peter] Well, it is still worthwhile to learn how (or even if) the device will work with the 2.4 kernel driver, even if you can't listen to music while you're doing the investigation :) But sounds like it doesn't work in 2.4 as you say, so easiest to debug what's going on using 2.6 for now.
(!) [Peter] Could you try instead:
modprobe gamecon map=0,1
Looking in ./drivers/input/joystick/gamecon.c it seems it needs the parallel port number and the pad number. I'm guessing that would be counting from 1, but maybe it should be map=0,0
module_param_array_named(map, gc, int, &gc_nargs, 0);
MODULE_PARM_DESC(map, "Describers first set of devices

(?) Solved!

First I had to remove lp module. I also stopped CUPS services because my printer is connected via parallel port.

(!) [Peter] Thanks, I didn't think of stopping CUPS :)

(?) In 2.4 kernel, I try:

#modprobe gamecon gc=0,7

where 0 means parallel port 0, and 7 means that it is a Playstation joystick.

(!) [Peter] the "map=0,7" would be what to use for 2.6, judging from the sources they mention that the "gc=" parameter is deprecated.

(?) When I tried to play xgalaga using my joystick, the ships did some strange movement.

(!) [Peter] Instead of tying a game directly, could you try to build the "jstest" and "jcal" programs if you don't have them already. From the README in the download mentioned before:


2.5 Verifying that it works
For testing the joystick driver functionality, there is the jstest program. You run it by typing:
       jstest /dev/js0
And it should show a line with the joystick values, which update as you move the stick, and press its buttons. The axes should all be zero when the joystick is in the center position. They should not jitter by themselves to other close values, and they also should be steady in any other position of the stick. They should have the full range from -32767 to 32767. If all this is met, then it's all fine, and you can play the games. :)
If it's not, then there might be a problem. Try to calibrate the joystick, and if it still doesn't work, read the drivers section of this file, the troubleshooting section, and the FAQ.
2.6. Calibration
For most joysticks you won't need any manual calibration, since the joystick should be autocalibrated by the driver automagically. However, with some analog joysticks, that either do not use linear resistors, or if you want better precision, you can use the jscal program
       jscal -c /dev/js0


(?) Maybe it still need some configuration or it's a bug in the kernel (the documentation in 2.4 kernel says that support for Playstation joystick is still under development).

(!) [Peter] Well, to debug we'd need a better description of the symptoms than "ships did some strange movement in xgalaga", I mean any kernel developer would probably find that quite insuffuicient :)

(?) But I'm quite happy to see that my Linux box is now more fun with joystick support :)

I haven't had time to try this on 2.6 kernel. Anyway thanks very much for the help :)

(!) [Peter] Could be that in 2.6 the driver would read more frequently from the parport than in 2.4, resulting in a more responsive joystick, so it could be worth a try. Especially to get sound to work at the same time, as you mentioned.

(?) LN -- have I gone brain dead?

From Bob van der Poel

Answered By: Breen Mullins, Thomas Adam, Ben Okopnik

I'm trying to do something pretty simple, I thought, and am near the "tear out my hair" stage ... All I want to do is to create a symbolic link to an existing directory:

bob$ ln -s tmp foo
bob$ ls -l foo
lrwxrwxrwx  1 bob bob 3 Dec  1 15:21 foo -> tmp/

No, foo is not a link to tmp. Well, it sort of is

bob$ file foo
foo: symbolic link to `tmp'


bob$ ls foo
(!) [Breen] It looks like you're somehow passing -F -d to the ls command. What does 'alias ls' tell you?

(?) Oh, that was easy! Here I've been struggling trying to get the *&*(^ ln command to work and it was working all along! Yup, ls was aliased:

alias ls
alias ls='ls -F --color=auto'

Guess how quickly I'm going to fix that! Dumb, dumb, dumb. Don't know who .. the packager who snuck in the -F or me :)

(!) [Breen] Moral: always check for aliases.
(!) [Ben] For some of us - those who have their "mv", "cp", and "rm" aliased to automatically use "-i", for example - it's the automatic response (sticking, say, 'rm' into a 'for' loop doesn't work too well if you're going to get asked "Are you sure?" for a couple hundred files.) However, there's other stuff of this ilk in shells that can really drive you nuts. In the past, I've been tripped up by looking for programs that didn't exist (they turned out to be functions that the sysadmin put in /etc/profile), have been driven almost insane by CDPATH ('cd foo' would go to '/usr/local/foo/bar/zotz/qux/argle/bargle/will/this/never/end/foo' instead of ./foo), and have become Quite Annoyed by per-directory source files - although '.exrc' isn't really a shell-related gadget. But still. :)
The Daemons of Unix are wily, subtle, and enjoy sucking out your brains through your ears if you let them. They're also a very strange shade of mauve with green dots, so avoid the psychedelic drugs and all will be well. Intoxicants and computing do NOT mix well... remember, never drink and derive.
(!) [Breen] Fedora Core 3 (in a fit of complete brain-death) shipped with vi unconditionally aliased to /usr/bin/vim.
(!) [Ben] Gaah! I've run into similar idiot's handiwork on a Solaris system - although, to be fair, it was the local sysadmin who perpetrated this one. He had set root's $SHELL to "/usr/bin/bash" - 'cause, y'know, "sh" is just so annoying (and .profile is just some stupid thing that nobody ever uses anyway, right?) The, one fine morning, the "/usr" slice failed to mount...
(!) [Breen] I've just taken another look at this -- at some point Fedora fixed the bug. /etc/profile.d/vim.sh now looks like this:
if [ -n "$BASH_VERSION" -o -n "$KSH_VERSION" -o -n "$ZSH_VERSION" ];
 [ -x /usr/bin/id ] || return
 [ `/usr/bin/id -u` -le 100 ] && return
 # for bash and zsh, only if no alias is already set
 alias vi >/dev/null 2>&1 || alias vi=vim
which is much more reasonable behavior. As originally shipped the two tests after then were missing.
(!) [Breen] Which doesn't matter much until you're trying to repair a system which doesn't have /usr/ mounted (because you borked /etc/fstab when tweaking your partitioning scheme).
Remembering to type /bin/vi is easier when you're not sweating bullets.

(?) doesn't give me a listing of the contents of tmp. Ummm, what am I doing wrong here?????

(!) [Thomas] Nothing. Although you should qualify which directory you created the link in. Indeed, what does:
ls -l ./foo
produce as output? Indeed, did you do that the right way around?
% ln -s /tmp foo && ls ./foo

(?) System V or BSD?

From Adam S Engel

Answered By: Thomas Adam, Rick Moen, Mike Orr

I was hunting down a "fugitive" process and accidently hit the BSD command ps aux (thinking it was an accident because I'm using Mandriva, which I thought was based on System V)

(!) [Thomas] In terms of Init level structures then it is, yes.

(?) and received something like this:

~ 495 --> ps aux | head -5
root         1  0.0  0.0   1560   476 ?        S    Nov30   0:00 init [5]
root         2  0.0  0.0      0     0 ?        SN   Nov30   0:00
root         3  0.0  0.0      0     0 ?        S<   Nov30   0:00 [events/0]
root         4  0.0  0.0      0     0 ?        S<   Nov30   0:00 [khelper]

But then, just for the hell of it, I hit the System V command, ps -ef and got this:

~ 494 --> ps -ef | head -5
root         1     0  0 Nov30 ?        00:00:00 init [5]
root         2     1  0 Nov30 ?        00:00:00 [ksoftirqd/0]
root         3     1  0 Nov30 ?        00:00:00 [events/0]
root         4     1  0 Nov30 ?        00:00:00 [khelper]

I thought most Linuxes, particularly the "big ones" like Fedore, Debian, SuSE, Mandriva, were based on System V.

(!) [Thomas] They are, but the ps command is an oddity in that sense. Still useful, though.

(?) Even so, my MacOSX, based loosely on BSD, will not accept the ps -ef command. Can a system (Mandriva in this case) be comprised of both?

(!) [Rick] All modern Unixes (except *BSD, which are holdouts) are the result of (or inspired by) a shotgun marriage of those two cultures (plus SunOS), called System V R4, released by AT&T in 1988. Quoting from the Unix FAQ:
Even prior to that, retrofitting "BSD enhancements" onto System V-based systems was a nearly ubiquitous customisation: Sys V R4 simply acknowledged that reality, bowing to the technical community's mandate.
The general architectural description of Sys V R4 got abstracted out as the POSIX.1 specification, which served as the blueprint from much of what Torvalds and co. (and FSF, and others) then performed in building up what today we call GNU/Linux systems: Linux distributions generally are an independent implementation of (most of) the POSIX spec, which in turn was based largely on Sys V R4. And so, here we are.

(?) The rc.init files seem, from what I recall from Redhat, to be System V.

(!) [Sluggo] See "man ps". The 'ps' Linux uses has two sets of command-line options to please everybody. There are also verbose long options, and two sets of output formats.
Linux generally follows the System V "style" but deviates whenever it feels like it.
(!) [Rick] And System V wasn't exactly System V, any more, after SysV R3. ;-> SysV R4 was, after all, one of computing biggest exercises in syncretism.[1]
(!) [Jimmy] The footnote is missing, but I assume it was to contain a definition of syncretism:
Attempted union of principles or parties irreconcilably at variance with each other.
(!) [Sluggo] The network stack is closer to BSD's.
(!) [Rick] Amen to that. Back when there was a Caldera Systems, Inc. (remember them?), they kept trying to get Torvalds to accept a huge patch to add the System V "STREAMS" interproces I/O system. Torvalds turned them down flat, repeatedly, saying (and I paraphrase): "Why would we need that bloatware in our kernel, when Berkeley sockets already do that same job at least as well?"
(!) [Sluggo] The init system is not part of "Linux" (the kernel) but is chosen by each distribution. Most distros use System V style, and even those that don't like Slackware have compatibility directories so that vendor-compiled commercial programs can be installed without too much hassle. However, the init strategy is one of the most varying things between distributions.

(?) I just got a new Dell Dimension basic "cheap" no-frills computer, but still, it's 80 gigs of hard-drive space are far more than i need, so I'm installing SuSE, Mandriva, and a third. It was going to be Debian, but now I'm thinking of FreeBSD. Can anyone tell me the major difference between FreeBSD and Linux?

(!) [Rick] Speaking in very broad terms, FreeBSD was designed and is maintained by people who never liked AT&T Unix or any of its derivatives and independent reimplementations at all, and who thus were unsatisfied by the System V R4 compromise. Of course, that's hardly all there is to it: If you'd been sued, more or less baselessly, by a Fortune 500 corporation for copyright infringement, you'd probably develop an attitude problem, too.
The difference in init scripts, you know about. A minority of Linux distributions have offered BSD init scripts, too, all along, e.g., Slackware, for those who like them.
Additionally, there are such a large number of other small but significant architectural differences that I cannot hope to remember them all.
The BSDs tend to favour a "BSD slice"-type partition map, as opposed to the IBM/Microsoft-style partition map more typical of Linux systems (at least on IA32/x86-64). Their directory layout is a bit different, not very close to the Filesystem Hierarchy Standard promoted by the Linux community. They tend to use variants on UFS as their main filesystem. (FreeBSD favours a variant called FFS = Fast Filesystem, incorporating Marshall Kirk McKusick's softupdates instead of journaling. Softupdates are a clever and effective way to get most of journaling's advantages in protecting metadata following crashes or power loss -- but sadly does nothing about the problem of long fsck times after such events.)
They tend not to have anything resembling /proc. They favour csh over Bourne variants such as bash. They favour nvi over vim. They default to use of the wheel group (which Linuxes usually lack mostly because Stallman at FSF doesn't like the concept). In general, their prejudices tend to come across to Linux users as a bit old-fogey-ish.
They offer both binary packages plus a very successful build-from-source architecture called the "ports" system. The closest thing to the latter on Linux would be Gentoo's portage and the GAR build system (and similar things in other build-from-source Linux distros).
They have their own libc (not based on GNU's, as Linux always has been) and their own implementation of make. Otherwise, a lot of the userspace toolsets are literally exactly the same (except compiled to a different ELF binary format).
Spend some time reading "The FreeBSD Handbook", online: It's a landmark in clarity for technical documentation, and worth reading on that basis alone.

(?) I've heard that FreeBSD is known for its security and other network capabilities.

(!) [Rick] FreeBSD is not especially known for security. (You might be thinking of OpenBSD.) I would say that it best known for scalable network performance under heavy load, and for stability. It's also known for having relatively narrow scope of hardware compatibility (i.e., driver selection), compared to Linux, and for relatively slow and conservative development: You may or may not find FreeBSD 6.0-STABLE's driver support for your new Dell Dimension's chipsets to be adequate.
Specific comparisons of FreeBSD with typical Linux systems have been done repeatedly, but tend to have the problem of being out of date (which doesn't stop people from alluding to them, out of ignorance). E.g., often-heard claims about better FreeBSD VM performance rely on Linux 2.4.x results, back when Linux had a real problem, there. For that matter, Linux's entire TCP/IP stack has been thrown out and rewritten four or five times during the kernel's history, so you can imagine the large opportunity for outdated comparisons, there.
It used to be the case that FreeBSD's SCSI subsystem gave significantly better performance than that of the Linux kernels, but I suspect that difference, too, has been eliminated. What probably does remain is the matter of NFS/autofs/amd: If you're a heavy NFS user, you'll find that FreeBSD's NFSv3 implementation still has fewer bugs than Linux's, though the gap is pretty small by now.
(!) [Sluggo] Some people say BSD's network stack can handle a higher load.
(!) [Rick] That's largely on the basis of data like this 1999 event: http://www.bafug.org/news/NewRecord.html
At that time, Walnut Creek CD-ROM set the world record for most bytes of network traffic processed in 24 hours by a single host: One single-processor PII box (a then-famous FreeBSD ftp server) handled 1.39 terabytes. (This burst of traffic was, ironically, occasioned by the release on that machine of Slackware 4.0.)
Fast-forward: About a year ago, I helped build what became the second-fastest computational cluster in the world, "Thunder", a set of 1024 quad-Itanium2 Linux boxes. The interconnects for node-to-node data passing used Quadrics cards, and I don't have figures handy but it was a stupendous level of network traffic.
(!) [Sluggo] But both systems are widely used in high-performance mission-critical situations, so the difference is a bit academic. Isn't Linux being used for rocket control somewhere?
Linux comes in a wide variety of flavors from user-compiled hacker systems (Gentoo) to turnkey Windows clones (Linspire), with RPM/DEB distros in between. BSD sticks to the user-compiled route.
(!) [Rick] Misconception. It's actually probably more common, in practice, to install software from the binary packages than to build it from ports. I can tell you that BSD admins will often resource to packages when there are local build problems, if nothing else.
(!) [Sluggo] BSD also tends to have less support for new hardware, and last time I checked it had an incompatible disk-partitioning scheme, making dual booting a problem.
(!) [Rick] FYI: FreeBSD (at least) is perfectly happy booting from an IBM/Microsoft-style partition table. For that matter, you can boot Linux (but not MS-Windows) from a BSD "slice"-style disklabel.
(!) [Sluggo] But if a user-compiled distro appeals to you and you like the BSD traditions, FreeBSD would be worth looking into. That seems to be the most popular one for new users, although NetBSD and OpenBSD are also free and would be worth comparing.
(!) [Rick] Don't overlook DragonFly BSD. http://linuxmafia.com/pipermail/conspire/2005-October/001504.html

(?) Missing libraries

...and how to find them

From Benjamin A. Okopnik

Answered By: Jimmy O'Regan, Neil Youngman, Peter Knaggs

All of a sudden, lots and lots of stuff - including 'vi' - is crashing when I try to bring it up. The error message I get is "error while loading shared libraries: libpangocairo-1.0.so.0: cannot open shared object file: No such file or directory".

Worst of all, grepping the Debian ls-lR doesn't show any such thing - and searching the Net has lots of people having the same problem and not being able to find a package that contains it. This is not sounding good.

Folks, if any of you could take a look in your /usr/lib (that's where 'strace' tells me these progs are looking for it) and send me a copy of your libpangocairo-1.0.so.0 - assuming that somebody somewhere has it - I'd be very grateful. Meanwhile, I'm quite annoyed and puzzled - how the heck can so much stuff depend on a lib that's not available???

Sigh. I hope I don't end up having to reinstall my entire system. That would be a really, really big problem while I'm on the road.

(?) Update: I've found an RPM that contains libpangocairo - presumably, it's something near what I need. Converting it wasn't useful, since it was going to put the files into a different directory - so I just copied out the files and put'em into /usr/lib.

Result: well, I've got Vim, Mozilla, and Firefox back. On the other hand, "mdh" (my MailDoHickey from Freshmeat that I've been using for a year or more) segfaults; so does "gqview".

My best guess as to the cause of this: earlier, I did an "apt-get update" and "apt-get dist-upgrade", and I recall seeing "libc" (and a few other libs) go flying by in the list of installed packages. If that's what it is, then I'm a bit shocked: I've never had Debian break my install before, simply via an update.

More tomorrow, since it's almost 2a.m. here.

(!) [Jimmy] New version of gvim? Pango is Gnome's framework for internationalised text (bidi, strange fonts, etc.), Cairo is a vector drawing library (like DPS or Apple's Display PDF (Quartz?)). All text in Gtk is now rendered through Pango, so everything that depends on Gtk in any way is going to depend on it.
It doesn't seem to be in Debian yet.

(?) Ah. I see.

A few days ago, the maintainer of Jpilot got back to me about a bug that I'd filed, and asked me to recompile Jpilot from source with the latest libpisock library. However, Debian's "official" method of creating a package from source is this nightmarish chase of dependencies, all alike... and Gtk+, Cairo, Pango, and a few other things (all from pretty much the same place - the gtk.org FTP server.)

However, everything worked OK back then - including the new version of JPilot. Something in the recent update must be conflicting with the "cutting edge" libs.

It still isn't looking good. I've tested a few other GTK-based apps - gtkpool, gtksee, gtop - and they work, although gtop throws out a cryptic warning:

glibtop: glibtop_get_swap (): Client requested field mask 0001f, but
only have 00007.

I have no idea what all those libs may have overwritten. Fixing it is going to take some thought. :|

(!) [Jimmy] Um... specifically, the Cairo backend for Pango doesn't seem to be in Debian yet, though Pango is.

(?) So, now that I'm actually awake, and possess a functioning brain - in contrast to last night - I have a plan of attack that should let me get past all this bull with the grace of a matador (and without using OLE even once.)

1) Pick a bunch of GTK-based apps and run each one. Add those that fail to a list. Then:

for app in $list
       ldd `which $app` |perl -wlne's/^.* => (\S+) .*/$1/;/gtk|pango|cairo/&&print'>>list.txt
       sort -uo list.txt list.txt

while read n; do readlink $n; done < list.txt

This should give me a list of all the relevant libs - running this for 'mdh' and 'gtk-gnutella' already shows some promise - that may need to be reinstalled. It may require that I manually remove the offending lib (sometimes, installing the right version doesn't do anything unless the old lib is removed), but that shouldn't be too difficult; the list won't be all that long.

Running the above for 'mdh' and 'gtk-gnutella' shows:


I'm betting that one of those - libgtk-x11-2.0.so.0, anybody? - is the bugger that's busting my chops. More tests later; gotta run to work NOW.

(?) Update: got it fixed. Most likely.  :) At least, GTK apps now run without complaining.

I looked at the list that I got as a result, and ran "dlocate" over the "root name" bits (everything up to the first '.'); other than the obvious, most of the rest pointed to libglib2.0-0. I reinstalled it and removed the newer versions - i.e., the current lib names mostly look like "libpangoft2-1.0.so.0.801.1", while the newer (broken) ones look like "libpangoft2-1.0.so.0.1001.0" - and life is good again. Whew.

I'm going to be doing more testing - i.e., by removing "libpangocairo", which should not be getting pulled in by anything - but it seems all right now.

(!) [Neil] There was a neat tip on TAG a while back about the LD_DEBUG environment variable. I think it could be useful in identifying the exact problem.
neil ~ 15:08:12 501 > LD_DEBUG=help ls
Valid options for the LD_DEBUG environment variable are:

 libs        display library search paths
 reloc       display relocation processing
 files       display progress for input file
 symbols     display symbol table processing
 bindings    display information about symbol binding
 versions    display version dependencies
 all         all previous options combined
 statistics  display relocation statistics
 unused      determined unused DSOs
 help        display this help message and exit

To direct the debugging output into a file instead of standard output
a filename can be specified using the LD_DEBUG_OUTPUT environment variable.
neil ~ 15:50:43 502 >

(?) Oh, good one! I wish I'd remembered it. I used 'strace' to see what was going on; unfortunately, it didn't show enough detail to be of use. The above may well do that; I'll use it to do a little troubleshooting, just to make sure that this is resolved. Thanks, Neil!

(!) [Peter] Ulrich Drepper has a quite readable guide to writing shared libraries, he's been maintaining it for quite a while now and in Jan 2005 put out this: http://people.redhat.com/drepper/dsohowto.pdf It's probably more intended for folks actually writing shared libs in the first place, but it's a good one for debugging

This page edited and maintained by the Editors of Linux Gazette
Copyright © its authors, 2006
Published in issue 122 of Linux Gazette January 2006
HTML script maintained by Heather Stern of Starshine Technical Services, http://www.starshine.org/

Published in Issue 122 of Linux Gazette, January 2006

A New Windoze Notebook — Now What? Knoppix!

By Edgar Howell


For a brief period of time this office had been Windows-free, in theory at least. The combo printer requiring '95 had stopped using black ink and '95 had stopped booting, making both obsolete. And '98 on the old notebook doesn't shutdown properly, which turns booting into an adventure. Just hadn't quite got around to dumping all this junk.

Then fate sent a project my way that required some variety of Windows. The one tool I absolutely need I got familiar with many years ago under OS/2. Unfortunately it no longer exists in any such incarnation. So it was time to clench my teeth, say good-bye to dreams of a notebook completely unexposed to Windows and order a Dell with whatever is current from Redmond now-a-days.

The Plan

We ordered the Dell with twice the hard-drive and 4 times the memory advertised, so we just reduce the partition a bit and put SuSE 10.0 on there and be up and running in nothing flat.

This is something I've done for years. One of my favorite commercial tools had been Partition Magic. Pretty easy to reduce a partition in size and use the free space for some other purpose.

But as Robert Burns put it: The best laid schemes o' Mice an' Men, Gang aft agley.

The Problem

If in the past I have ever said anything nice about products from Redmond, I really do regret it. If anything, I have certainly been far too polite in just using words like "junk". There are other 4-letter and longer words far more suitable.

Consider: you buy a machine with one pre-installed operating system on an 80GB drive and cfdisk reports:

              First       Last
 # Type       Sector      Sector   Offset    Length  File System Type (ID) Flags
-- ------- ----------- ----------- ------ ----------- -------------------- -----
 1 Primary           0      176714     63      176715 Dell Utility (DE)    none
 2 Primary      176715   146544929      0   146368215 HPFS/NTFS (07)       Boot
   Pri/Log   146544930   146560994      0       16065 free                 none
 3 Primary   146560995   156296384      0     9735390 CP/M / CTOS / . (DB) none

Mount the 3 primary partitions and df shows:

Filesystem           1K-blocks      Used Available Use% Mounted on
/UNIONFS/dev/sda1        88136      6654     81482   8% /mnt/sda1
/UNIONFS/dev/sda2     73184104   7132524  66051580  10% /mnt/sda2
/UNIONFS/dev/sda3      4858184   3167052   1691132  66% /mnt/sda3

Why is there more than one primary partition? The hardware architecture only supports 4, they're valuable! None is particularly full. And that gap in the middle?! Can we spell "fragmentation"? Define it?

Maybe people living in the wild and wooly world of Windows are used to things like this and don't consider it fraudulent when deprived of about 14GB out of 80. But to have that much real estate taken away without even asking!? Consider the percentage if I hadn't chosen a drive twice the size of normal.

Is there any need to point out the bizarre file system types? And I don't do NTFS!

Tabula Rasa

But it gets better, or worse. Back in '95 days, you got a CD with the operating system, probably also pre-installed on the machine.

With '98 the system was pre-installed and you got a so-called "Recovery CD" along with it. At some point in time you were forced to use the infernal thing and that was when you learned that it wouldn't just restore the operating system but would return the entire partition to the status it had originally had when the machine left the manufacturer.

Like good-bye data if you don't have adequate back-ups and/or another bootable partition (this was long before Knoppix). But if you modified the partition size in order to install something else as well, the "Recovery CD" fails — after it has formatted the partition it can't use! And the last Partition Magic I bought couldn't help because instead of running under DOS it required Windows. A perverse Catch-22. Been there, done that, wore out the T-shirt.

But at least you had a CD with the system software on it.

Dell/Microsoft didn't even provide one! Just how expensive are CD-ROMs today? In bulk?

That's what those other partitions are all about. In other words it is impossible to do something simple like format the drive, allocate partitions, and install software. So now what?!

Tools and Toolboxes

In spite of having used other fine tools in the past to work through problems, Knoppix has become my toolbox of choice because it gives me an environment I am familiar with, GNU/Linux, and lots and lots of tools.

Of course the first thing I did — after minimal setup, as little as possible — was to change the boot sequence on the notebook to enable using Knoppix. I had to know how to do this (F2, DEL early in the boot process are good candidates): other than one large piece of paper (roughly 2-times legal-size) describing the external features of the hardware, Dell included no significant printed documentation. It's all on the hard-drive.

Above, you have already seen some of the information available from use of the tools on the Knoppix CD/DVD. By the way, here I am using a DVD with Knoppix 4.0.

So far all we have been able to do is to confirm that we have a major problem. Bizarre HD configuration. Unfamiliar file system.

So let's see what tools Klaus Knopper included in his toolbox:

knoppix@3[knoppix]$ apropos NTFS
libntfs-gnomevfs (8) - Module for GNOME VFS that allows access to NTFS filesystems.
mkntfs (8)           - create a NTFS 1.2 (Windows NT/2000/XP) file system
ntfscat (8)          - concatenate files and print them on the standard output
ntfsclone (8)        - Efficiently clone, image or restore an NTFS filesystem
ntfscluster (8)      - identify files in a specified region of an NTFS volume.
ntfsfix (8)          - tool for fixing NTFS partitions altered by the Linux kernel NTFS driver.
ntfsinfo (8)         - dump a file's attributes
ntfslabel (8)        - display/change the label on an ntfs file system
ntfsls (8)           - list directory contents on an NTFS filesystem
ntfsprogs (8)        - several tools for doing neat things with NTFS partitions
ntfsresize (8)       - resize an NTFS filesystem without data loss
ntfsundelete (8)     - recover a deleted file from an NTFS volume.
smbcquotas (1)       - Set or get QUOTAs of NTFS 5 shares

Plan B

From looking at a couple of the man pages it would seem that this collection of tools has been around for 10 years or so, well beyond release 0.0 by now.

So let's use ntfsclone to push an image of the bootable partition over the LAN onto another machine. Once we have that as back-up we can use ntfsresize to shrink the partition down to something reasonable. After that it should be easy enough to create an extended partition with a couple of logical partitions for Linux.

I'm not at all familiar with these tools but it is always fun to learn something new. And the time is right.

Backing Up

Again, nothing ever quite works the way one expects. Making the image was no exception. Here is one of the unsuccessful attempts:

knoppix@3[knoppix]$ su
root@3[knoppix]# ntfsclone --save-image --output - /dev/sda2 | gzip -c | \
ssh -l web lohgopc2 'cat > /DATA/NO_BACKUP/DELL/sda2.img.gz'
ntfsclone v1.9.4
The authenticity of host 'lohgopc2 (' can't be established.
RSA key fingerprint is 65:32:cc:81:8f:eb:73:24:7b:b3:18:a8:66:fa:7c:ae.
Are you sure you want to continue connecting (yes/no)? NTFS volume version: 3.1
Cluster size       : 4096 bytes
Current volume size: 74940522496 bytes (74941 MB)
Current device size: 74940526080 bytes (74941 MB)
Scanning volume ...
100.00 percent completed
Accounting clusters ...
Space in use       : 7304 MB (9,7%)
Saving NTFS to image ...


After "Saving NTFS to image ..." nothing further happened and I had to cancel with CTRL-C.

Here is what finally worked:

knoppix@1[knoppix]$ su
sudo: unable to lookup Knoppix via gethostbyname()
root@1[knoppix]# ntfsclone --save-image --output - /dev/sda2 | gzip -c | \
ssh -l web lohgopc2 'cat > /DATA/NO_BACKUP/DELL/sda2.img.gz'
ntfsclone v1.9.4
Password: NTFS volume version: 3.1
Cluster size       : 4096 bytes
Current volume size: 74940522496 bytes (74941 MB)
Current device size: 74940526080 bytes (74941 MB)
Scanning volume ...
100.00 percent completed
Accounting clusters ...
Space in use       : 7304 MB (9,7%)
Saving NTFS to image ...

Warning: No xauth data; using fake authentication data for X11 forwarding.
100.00 percent completed
Syncing ...

Other than the obvious difference between success and failure, from a quick glance at the output it isn't clear why one worked and one didn't. But look at the lines ending in "NTFS volume version: 3.1".

On the unsuccessful attempt I had failed to note that ssh wanted confirmation that it was OK to establish a connection with the other machine. It was patiently waiting for me to enter "yes" while I was scratching my head trying to figure out why nothing was happening!

Success the next morning — a good night's sleep can do a world of good — didn't include that question because I had already told ssh that it was OK and it had "permanently" (until next boot with Knoppix) recorded this fact. Refreshed and thinking clearly, I remembered that the password was needed and saw the request from ssh.


Now that our safety-net was in place the first thing to do was to check out the partition:

knoppix@4[knoppix]$ sudo ntfsresize --info /dev/sda2
sudo: unable to lookup Knoppix via gethostbyname()
ntfsresize v1.9.4
NTFS volume version: 3.1
Cluster size       : 4096 bytes
Current volume size: 74940523008 bytes (74941 MB)
Current device size: 74940526080 bytes (74941 MB)
Checking filesystem consistency ...
100.00 percent completed
Accounting clusters ...
Space in use       : 7304 MB (9,7%)
Collecting shrinkage constrains ...
Estimating smallest shrunken size supported ...
File feature         Last used at      By inode
$MFT               :     16790 MB             0
Multi-Record       :      8393 MB             9
You might resize at 7303708672 bytes or 7304 MB (freeing 67637 MB).
Please make a test run using both the -n and -s options before real resizing!

Rather than reducing the partition as far as possible, let's follow that advice and see what it looks like if we leave some room for data and the tool that has yet to be installed:

root@0[knoppix]# ntfsresize --no-action --size 10G /dev/sda2
ntfsresize v1.9.4
NTFS volume version: 3.1
Cluster size       : 4096 bytes
Current volume size: 74940523008 bytes (74941 MB)
Current device size: 74940526080 bytes (74941 MB)
New volume size    : 9999995392 bytes (10000 MB)
Checking filesystem consistency ...
100.00 percent completed
Accounting clusters ...
Space in use       : 7304 MB (9,7%)
Collecting shrinkage constrains ...
Needed relocations : 395906 (1622 MB)
Schedule chkdsk for NTFS consistency check at Windows boot time ...
Resetting $LogFile ... (this might take a while)
Relocating needed data ...
100.00 percent completed
Updating $BadClust file ...
Updating $Bitmap file ...
Updating Boot record ...
The read-only test run ended successfully.

OK, go for it.

root@0[knoppix]# ntfsresize --size 10G /dev/sda2
ntfsresize v1.9.4
NTFS volume version: 3.1
Cluster size       : 4096 bytes
Current volume size: 74940523008 bytes (74941 MB)
Current device size: 74940526080 bytes (74941 MB)
New volume size    : 9999995392 bytes (10000 MB)
Checking filesystem consistency ...
100.00 percent completed
Accounting clusters ...
Space in use       : 7304 MB (9,7%)
Collecting shrinkage constrains ...
Needed relocations : 395906 (1622 MB)
WARNING: Every sanity check passed and only the DANGEROUS operations left.
Please make sure all your important data had been backed up in case of an
unexpected failure!
Are you sure you want to proceed (y/[n])? y
Schedule chkdsk for NTFS consistency check at Windows boot time ...
Resetting $LogFile ... (this might take a while)
Relocating needed data ...
100.00 percent completed
Updating $BadClust file ...
Updating $Bitmap file ...
Updating Boot record ...
Syncing device ...
Successfully resized NTFS on device '/dev/sda2'.
You can go on to shrink the device e.g. with 'fdisk'.
IMPORTANT: When recreating the partition, make sure you
  1)  create it with the same starting disk cylinder
  2)  create it with the same partition type (usually 7, HPFS/NTFS)
  3)  do not make it smaller than the new NTFS filesystem size
  4)  set the bootable flag for the partition if it existed before
Otherwise you may lose your data or can't boot your computer from the disk!

Now that we have reduced the size of the area used within the physical partition (comparable to de-frag under Windows, maybe) it is safe to use cfdisk /dev/sda

                                  Festplatte: /dev/sda
                            Größe: 80026361856 Bytes, 80,0 GB
                   Köpfe: 255   Sektoren pro Spur: 63   Zylinder: 9729

    Name         Flags        Part. Typ   Dateisystemtyp     [Bezeichner]   Größe (MB)
    sda1                       Primäre    Dell Utility                           90,48
    sda2         Boot          Primäre    NTFS               []               10001,95
    sda5                       Logische   Linux ext2                          10010,17
    sda6                       Logische   Linux                               10001,95
                               Logische   Freier Bereich                      14928,89
    sda7                       Logische   FAT16                               10001,95
    sda8                       Logische   Linux swap / Solaris                20003,89
    sda3                       Primäre    CP/M / CTOS / ...                    4984,52

     [ Bootbar]  [Löschen ]  [ Hilfe  ]  [ Maxim. ]  [Ausgabe ]  [  Ende  ]
     [  Typ   ]  [Einheit.]  [Schreib.]

This was the step where extreme care was called for. A mistake with ntfsresize produces an error message or warning. To me the partition table is comparable to the base register in some assembler languages: something that cannot be verified, a promise, not a guarantee. Mess it up and you may never be able to recover. Pay very close attention to the changes you make and back out if there is anything at all that you don't fully understand!

Finishing Touches

Once again, in the white hat, Linux wins! The backup was unneeded, although I never would have continued without one.

When booted into Windows, it did whatever it is that it does while it is doing what it does when one doesn't know what it is doing. So that worked. I guess.

Creating the extended partition and a couple of logical partitions inside it is quite straight-forward and doesn't deserve further discussion here. Although I grew up with fdisk, I will repeat the recommendation in the documentation to use cfdisk instead. Interactive and keeping one apprised of current status, it was very easy to get used to.


In retrospect it wasn't necessary to jump back and forth between normal user and root as in the cut-and-paste from screens, above. But there is a time and place for each. On more than one occasion I have failed to notice that I had returned to a virtual terminal with root privileges. Bad! The preferred idiom with Knoppix is "sudo". You won't need a root password but it keeps you aware of where you are and what you are doing.

After having gone through this exercise it occurred to me that a better long-term solution might be to back up the Windows partition as described after major changes, i.e. installation of the tool I need. The other two as well of course, just in case something goes wrong. And then when the time comes that the Windows partition becomes unusable and restoring is necessary, just do it from the network. That should make those other two partitions obsolete and recover a bunch of disk space. Way down the road.

The modem is presumably one of those nefarious "winmodems". I didn't check it out, ISDN currently, not analog. The network card works without problem. And the Kensington Mouse at the USB port was recognized by Knoppix off the starting blocks. It is really neat, the cable winds up into it, to the length needed, from 0 to whatever. I may even take it along when on the road.

The Dell hardware is actually very nice. If you could just get it without having to pay for an operating system you normally don't need...

[BIO] Edgar is a consultant in the Cologne/Bonn area in Germany. His day job involves helping a customer with payroll, maintaining ancient IBM Assembler programs, some occasional COBOL, and otherwise using QMF, PL/1 and DB/2 under MVS.

(Note: mail that does not contain "linuxgazette" in the subject will be rejected.)

Copyright © 2006, Edgar Howell. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 122 of Linux Gazette, January 2006

Benchmarking Filesystems Part II

By Justin Piszcz


After the last article was published, I have received more than a dozen requests for a second filesystem benchmark using the 2.6 kernel. Since that time, I have converted entirely to XFS for every Linux machine I use, so I may be a bit bias regarding the XFS filesystem. I tried to keep the hardware roughly the same. Instead of a Western Digital 250GB and Promise ATA/100 controller, I am now am using a Seagate 400GB and Maxtor ATA/133 Promise controller. The physical machine remains the same, there is an additional 664MB of swap and I am now running Debian Etch. In the previous article, I was running Slackware 9.1 with custom compiled filesystem utilities. I've added a small section in the beginning that shows the filesystem creation and mount time, I've also added a graph showing these new benchmarks. After the first round of benchmarks, I received a sleuth of e-mails asking for the raw numbers. The numbers are now included in tables at the end of this e-mail for both the last and current set of benchmarks.

What's new?

1) Unify the graphs so all but one are the same type.
2) Run tests with a recent distribution and the kernel.
3) Included the ReiserFS Version 4 benchmarks.
4) Included the raw data in matrix form at the bottom of this page.
5) Included three additional graphs:
   a) Creation time to make the actual filesystem.
   b) Time it takes to mount the filesystem.
   c) The amount of space available after formatting with the default options.



    COMPUTER: Dell Optiplex GX1
         CPU: Pentium III 500MHZ
         RAM:  768MB
        SWAP: 2200MB
  CONTROLLER: Maxtor Promise ATA/133 TX2 - IN PCI SLOT #1 
 DRIVES USED: 1] Seagate 400GB ATA/100 8MB CACHE 7200RPM
              2] Maxtor 61.4GB ATA/66 2MB CACHE 5400RPM
DRIVE TESTED: The Seagate 400GB.


       KERNEL: linux-
COMPILER USED: gcc-4.0.3
         EXT2: e2fsprogs-1.38/sbin/mkfs.EXT2
         EXT3: e2fsprogs-1.38/sbin/mkfs.EXT3
          JFS: jfsutils-1.1.8/sbin/mkfs.jfs 
   REISERFSv3: reiserfsprogs-3.6.19/sbin/mkreiserfs
   REISERFSv4: reiser4progs-1.0.5/sbin/
               (Used patch reiser4-for-2.6.14-1.patch w/ libaal-1.0.5 + reiser4progs-1.0.5)
          XFS: xfsprogs-2.6.36/sbin/mkfs.xfs

Tests Performed

001] Create 10,000 files with touch in a directory.
002] Run 'find' on that directory.
003] Remove the directory.
004] Create 10,000 directories with mkdir in a directory.
005] Run 'find' on that directory.
006] Remove the directory containing the 10,000 directories.
007] Copy kernel tarball from other disk to test disk.
008] Copy kernel tarball from test disk to other disk.
009] Untar kernel tarball on the same disk.
010] Tar kernel tarball on the same disk.
011] Remove kernel source tree.
012] Copy kernel tarball 10 times.
013] Create 1GB file from /dev/zero.
014] Copy the 1GB file on the same disk.
015] Split a 10MB file into 1000/1024/2048/4096/8192 byte pieces.
016] Copy kernel source tree on the same disk.
017] Cat a 1GB file to /dev/null.

NOTE1: Between each test run, a 'sync' and 10 second sleep   
       were performed.
NOTE2: Each file system was tested on a cleanly made file
NOTE3: All file systems were created using default options.
NOTE4: All tests were performed with the cron daemon killed   
       and with 1 user logged in.
NOTE5: All tests were run 3 times and the average was taken,  
       if any tests were questionable, they were re-run and 
       checked with the previous average for consistency.

Creating the Filesystems


p500:~# mkfs.EXT2 /dev/hde1
mke2fs 1.38 (30-Jun-2005)
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
48840704 inodes, 97677200 blocks
4883860 blocks (5.00%) reserved for the super user
First data block=0
2981 block groups
32768 blocks per group, 32768 fragments per group
16384 inodes per group
Superblock backups stored on blocks: 
        32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 
        4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968

Writing inode tables: done                            
Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 34 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.


p500:~# mkfs.EXT3 /dev/hde1
mke2fs 1.38 (30-Jun-2005)
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
48840704 inodes, 97677200 blocks
4883860 blocks (5.00%) reserved for the super user
First data block=0
2981 block groups
32768 blocks per group, 32768 fragments per group
16384 inodes per group
Superblock backups stored on blocks:
        32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
        4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968

Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 34 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.


p500:~# mkfs.jfs -q /dev/hde1
mkfs.jfs version 1.1.8, 03-May-2005

Format completed successfully.

390708801 kilobytes total disk space.
0.03user 1.01system 0:02.62elapsed 40%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+224minor)pagefaults 0swaps


p500:~# mkreiserfs -q /dev/hde1
mkreiserfs 3.6.19 (2003 www.namesys.com)

A pair of credits:
Continuing core development of ReiserFS is  mostly paid for by Hans Reiser from
money made selling licenses  in addition to the GPL to companies who don't want
it known that they use ReiserFS  as a foundation for their proprietary product.
And my lawyer asked 'People pay you money for this?'. Yup. Life is good. If you
buy ReiserFS, you can focus on your value add rather than reinventing an entire

Chris Mason wrote the journaling code for V3,  which was enormously more useful
to users than just waiting until  we could create a wandering log filesystem as
Hans would have unwisely done without him.
Jeff Mahoney optimized the bitmap  scanning code for V3,  and performed the big
endian cleanups. 


p500:~# mkfs.reiser4 -y /dev/hde1


p500:~# mkfs.xfs -f /dev/hde1
meta-data=/dev/hde1              isize=256    agcount=16, agsize=6104825 blks
         =                       sectsz=512  
data     =                       bsize=4096   blocks=97677200, imaxpct=25
         =                       sunit=0      swidth=0 blks, unwritten=1
naming   =version 2              bsize=4096  
log      =internal log           bsize=4096   blocks=32768, version=1
         =                       sectsz=512   sunit=0 blks
realtime =none                   extsz=65536  blocks=0, rtextents=0

Benchmark Set 1 of 4

Both extended filesystem revisions take a backseat to their counterparts.

ReiserFS takes a VERY long time to mount the filesystem. I included this test because I found it actually takes minutes to hours mounting a ReiserFS filesystem on a large RAID volume.

People always complain about how hard drive manufacturers equate 1000 kilobytes as a megabyte, well the filesystem is also part of the problem!

Benchmark Set 2 of 4

In the first test, ReiserFSv3 continues to lead the pack, with ReiserFSv4 and JFS not far behind.

ReiserFSv4 is now the slowest filesystem to search for files, previously, it had been XFS.

XFS is no longer the slowest filesystem for removing many files. However, EXT2 and EXT3 still win by far.

Similar to the first test, EXT2 and EXT3 lag behind the other journaling filesystems.

Contrary to the first set of benchmarks I ran, it appears XFS has slightly improved, ReiserFSv4 is now the slowest with ReiserFSv3 behind it. Also in the previous test, EXT3 had suffered a huge performance hit in this area, it is now comparable to EXT2.

It appears ReiserFS v3 and v4 have some serious issues deleting large numbers of directories.

In the first benchmark test, EXT2 won the test, now XFS is the new leader.

This benchmark represents how fast the tar ball can be read from each file system. Surprisingly, XFS now exceeds EXT3.

ReiserFSv3 won the last benchmarking round; however, EXT2 and EXT3 now dominate this test.

The best journaling file system here used to be ReiserFSv3; however, JFS now wins the benchmark.

ReiserFSv3 used to be the winner, it is now ousted by EXT2 and EXT3.

The biggest difference here is most of the filesystems have evened out compared to the first benchmarking test.

This test has remained relatively the same.

Once again, mostly the same; however, it is important to note that the performance of EXT2 and EXT3 are much closer this time.

EXT2 and EXT3 take the biggest hit up to about 4096 bytes which is surprising.

EXT2 and EXT3 now outperform ReiserFSv3. In most of these benchmarks thus far, ReiserFSv4 is the slowest; however, here we see that V4 is about 12 seconds faster than V3.

JFS continues to lead this test with EXT2 and EXT3 not far behind.

The following represents the combined test times.

Benchmark Set 3 of 4

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

The following is the CPU utilization for this test.

Here is a line chart of representing all of the test times.

Benchmark Set 4 of 4

Here is a composite of the total test time for all tests.

File Benchmark II Data

Seconds Elapsed
001 Touch 10,000 Files 48.25 48.25 34.59 33.59 34.08 37.47
002 Find 10,000 Files 0.03 0.03 0.03 0.03 0.07 0.04
003 Remove 10,000 Files 0.16 0.16 1.64 1.84 2.98 2.51
004 Make 10,000 Directories 49.76 49.87 34.32 33.74 34.68 37.17
005 Find 10,000 Directories 0.65 0.65 0.63 1.07 1.46 0.72
006 Remove 10,000 Directories 1.66 1.67 3.58 43.48 119.42 5.39
007 Copy Tarball from Other to Current Disk 5.17 5.15 5.74 5.12 7.34 4.26
008 Copy Tarball from Current to Other Disk 6.96 7.00 6.97 6.89 8.21 6.69
009 UnTAR Kernel Tarball 14.92 15.19 27.64 26.92 21.45 40.81
010 TAR Kernel Source Tree 14.05 14.08 13.05 33.49 25.82 36.19
011 Remove Kernel Source Tree 2.47 2.64 6.17 5.65 10.15 9.10
012 Copy Tarball 10 Times 39.48 38.29 39.13 45.15 62.16 46.34
013 Create a 1GB File 15.02 15.02 15.12 15.96 25.40 15.87
014 Copy a 1GB File 36.87 36.51 38.54 47.60 50.63 41.25
015 Split 10M File into 1000 Byte Pieces 57.26 57.77 2.99 4.35 2.95 4.87
016 Split 10M File into 1024 Byte Pieces 28.73 28.97 2.24 4.04 2.61 4.01
017 Split 10M File into 2048 Byte Pieces 7.02 6.98 1.39 2.26 1.55 1.95
018 Split 10M File into 4096 Byte Pieces 1.85 1.83 0.67 1.05 0.99 0.98
019 Split 10M File into 8192 Byte Pieces 0.58 0.58 0.36 0.56 0.62 0.57
020 Copy Kernel Source Tree 10.02 10.06 35.76 31.64 20.17 43.42
021 CAT 1GB File to /dev/null 18.90 18.59 18.00 37.33 21.37 18.70
CPU Utilization
001 Touch 10,000 Files 99.00 99.00 99.00 99.00 99.33 99.00
002 Find 10,000 Files 94.00 93.00 94.00 95.00 97.00 95.66
003 Remove 10,000 Files 98.00 98.66 73.66 99.00 99.00 91.66
004 Make 10,000 Directories 98.00 97.33 99.00 99.00 99.66 99.00
005 Find 10,000 Directories 99.00 99.00 99.00 99.00 99.00 99.00
006 Remove 10,000 Directories 99.00 99.00 88.66 99.00 99.00 97.00
007 Copy Tarball from Other to Current Disk 74.66 74.66 76.00 74.66 61.33 92.33
008 Copy Tarball from Current to Other Disk 60.00 59.33 59.33 62.00 86.00 62.66
009 UnTAR Kernel Tarball 42.33 41.33 27.33 53.00 80.00 26.00
010 TAR Kernel Source Tree 44.00 43.66 51.33 26.66 48.66 21.00
011 Remove Kernel Source Tree 39.66 36.66 33.00 89.33 88.33 63.66
012 Copy Tarball 10 Times 79.33 80.66 93.33 74.33 73.00 90.33
013 Create a 1GB File 56.00 55.66 67.33 57.00 50.00 64.33
014 Copy a 1GB File 42.00 42.00 47.00 37.33 52.00 49.33
015 Split 10M File into 1000 Byte Pieces 99.00 99.00 64.33 96.33 98.00 86.33
016 Split 10M File into 1024 Byte Pieces 99.00 99.00 77.33 97.66 99.00 97.00
017 Split 10M File into 2048 Byte Pieces 99.00 99.00 64.00 96.66 99.00 97.33
018 Split 10M File into 4096 Byte Pieces 99.00 99.00 69.33 99.00 99.00 97.33
019 Split 10M File into 8192 Byte Pieces 99.00 99.00 87.00 89.66 99.00 97.66
020 Copy Kernel Source Tree 65.33 65.00 21.33 41.33 70.33 25.33
021 CAT 1GB File to /dev/null 26.33 27.00 27.33 36.66 46.33 30.00

File Benchmark I Data


With the second round of filesystem benchmarks, I hope everyone is now satisfied with the benchmarks using the 2.6 kernel. What I gleam from these benchmarks is both EXT2 and EXT3 are now roughly the same speeds in the majority of the tests. It also appears the XFS has improved in the majority of the tests. ReiserFSv3 has slowed in many of the tests with ReiserFSv4 being the slowest in most of the tests. It is important to note that JFS has improved in some of the tests. Personally, I still choose XFS for filesystem performance and scalability.

Seconds Elapsed
Test Filesystem EXT2 EXT3 JFS REISER XFS
001 Touch 10,000 Files 34.14 40.98 24.50 23.03 27.13
002 Find 10,000 Files 0.02 0.02 0.02 0.02 0.04
003 Remove 10,000 Files 0.14 0.36 1.37 1.76 2.52
004 Make 10,000 Directories 73.53 148.46 24.11 23.42 29.19
005 Find 10,000 Directories 0.44 0.66 0.41 0.87 0.59
006 Remove 10,000 Directories 1.97 56.31 4.05 5.37 6.70
007 Copy Tarball from Other to Current Disk 3.26 5.70 4.62 6.08 4.12
008 Copy Tarball from Current to Other Disk 3.36 3.45 3.49 3.53 3.45
009 UnTAR Kernel 2.4.26 Tarball 24.49 31.73 34.64 12.36 23.79
010 TAR Kernel 2.4.26 Source Tree 17.86 23.40 27.06 22.81 24.85
011 Remove Kernel 2.4.26 Source Tree 4.24 7.26 10.86 3.18 4.48
012 Copy 2.4.26 Tarball 10 Times 18.28 46.68 38.17 49.16 26.22
013 Create a 1GB File 18.93 22.35 28.87 25.80 20.49
014 Copy a 1GB File 45.04 62.48 54.46 71.06 55.89
015 Split 10M File into 1000 Byte Pieces 53.35 62.70 8.11 4.15 4.34
016 Split 10M File into 1024 Byte Pieces 26.24 32.61 7.63 3.90 3.96
017 Split 10M File into 2048 Byte Pieces 6.40 7.94 2.83 2.22 1.95
018 Split 10M File into 4096 Byte Pieces 1.65 2.10 0.50 0.65 0.97
019 Split 10M File into 8192 byte Pieces 0.51 0.69 0.31 0.47 0.53
020 Copy 2.4.26 Kernel Source Tree 24.43 30.65 25.68 22.19 26.53
021 CAT 1GB File to /dev/null 21.70 23.52 20.40 23.28 21.13
CPU Utilization
Test Filesystem EXT2 EXT3 JFS REISER XFS
001 Touch 10,000 Files 97.66 96.66 92.66 97.66 97.00
002 Find 10,000 Files 98.33 107.33 109.66 85.66 80.66
003 Remove 10,000 Files 95.66 98.33 76.66 97.66 91.33
004 Make 10,000 Directories 63.00 33.00 96.00 97.66 90.66
005 Find 10,000 Directories 96.66 97.00 99.33 95.33 98.66
006 Remove 10,000 Directories 99.00 6.33 87.66 98.00 89.33
007 Copy Tarball from Other to Current Disk 74.66 79.00 79.33 90.33 81.66
008 Copy Tarball from Current to Other Disk 67.00 64.66 64.00 70.66 75.66
009 UnTAR Kernel 2.4.26 Tarball 16.00 18.66 15.33 78.33 31.66
010 TAR Kernel 2.4.26 Source Tree 21.66 22.66 17.00 32.00 18.33
011 Remove Kernel 2.4.26 Source Tree 11.00 18.33 9.00 96.66 77.33
012 Copy 2.4.26 Tarball 10 Times 96.66 67.33 66.33 96.00 92.00
013 Create a 1GB File 35.66 57.66 32.33 84.33 39.00
014 Copy a 1GB File 30.33 41.33 40.66 53.33 40.00
015 Split 10M File into 1000 Byte Pieces 97.66 96.66 22.33 89.33 92.33
016 Split 10M File into 1024 Byte Pieces 97.66 97.33 20.33 89.00 95.66
017 Split 10M File into 2048 Byte Pieces 95.33 97.33 26.00 88.66 95.66
018 Split 10M File into 4096 Byte Pieces 98.33 98.00 80.00 97.33 92.33
019 Split 10M File into 8192 Byte Pieces 99.33 98.33 87.00 96.00 94.66
020 Copy 2.4.26 Kernel Source Tree 19.33 17.33 18.00 44.33 25.00
021 CAT 1GB File to /dev/null 26.00 26.00 30.00 30.00 33.66

Copyright © 2006, Justin Piszcz. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 122 of Linux Gazette, January 2006

/dev/fanout : A One-To-Many Multiplexer

By Bob Smith

A one-to-many multiplexer Introduction
Build and Install /dev/fanout
How /dev/fanout Works
Security and Obsolescence


This article describes a Linux module that replicates its input on all of its outputs, a so called "fanout" or "one to many" multiplexer.

Purpose: The purpose of fanout is to give Linux a simple broadcast IPC.

Our own purpose for writing the module was to distribute log messages to one or more processes that want to be notified when an event occurs. We use /dev/fanout, a web server, and XMLHttpRequest on a web client to build an alarm system with multiple web interfaces running simultaneously. One nice feature of our alarm system is that the web interfaces don't use polling but still update automatically when a new alarm system message arrives.

Common Approaches to Broadcast: The two most common broadcast mechanisms in Linux are signals and UDP packets. You can broadcast a signal to a group of related processes using the kill command with a PID of zero. This works well if all of the processes are related and if the program knows what action is required on your signal.

Signals will not work for our application since there is no way to directly route a signal from a web server to a web client, and because web servers do not know that we want to redraw certain web screens on a particular signal.

We can also broadcast events using UDP or TCP. We've built event servers which accept TCP connections and broadcast event information down each accepted connection. We use XMLHttpRequest to request a PHP page that opens the TCP connection and waits for the event. While this approach works well, it requires yet another process and has the slight extra burden of an additional TCP connection for each web client.

A Better Broadcast Approach: A better approach would be to have something like a FIFO, but instead of having all of the listeners compete for the single copy of the input message, have all of the listeners get their own copy. Consider the following bash dialog:

  mkfifo event_fifo
  cat event_fifo &
  cat event_fifo &
  cat event_fifo &
  echo "Hello World" > event_fifo
  Hello World

The message appears only once, since only one instance of the cat command is given the fifo output. Now let's consider the same experiment using fanout:

  cat /dev/fanout &
  cat /dev/fanout &
  cat /dev/fanout &
  echo "Hello World" > /dev/fanout
  Hello World
  Hello World
  Hello World

The message now appears once for each of the three listening cat commands. We use bash commands just to illustrate what fanout does. Its real power lies in letting many different programs get identical copies of a data stream.

Which fails: send or receive? No matter how hard we try to avoid it, one day we'll find a reading process that can not keep up with the writing process. Allocating more memory postpones the problem but does not eliminate it. When this problem occurs we have two choices: apply back pressure to the writer causing the writer to block, or let the readers miss some output.

The problem with blocking the writing process is that you may affect other parts of the system. Our original purpose was to build an alarm system and we chose to route all event notifications through syslogd. Since we have /dev/fanout as a target in the syslog.conf file, blocking the writer would block syslogd and defeat the whole purpose of our alarm system.

The author of fanout very deliberately chose to cause the reader to fail when it can not keep up. Data is stored in a circular buffer and if a reader can not keep up with the writer, it will eventually ask for data that is no longer in the circular buffer. The fanout device returns an EPIPE error to the reader when this happens. In our application for /dev/fanout we are happy to protect syslogd at the expense of the web clients when we are forced to choose one over the other.

Build and Install /dev/fanout

The source code to the fanout module is available as a compressed tarball fanout.tgz, or you can pick up the individual files, fanout.c and Makefile.

Build the module with the following commands:

  cd /usr/src/linux
  tar -xzf fanout.tgz
  cd fanout

When you install the module you can set the size of the circular buffer and can set the verbosity of the printk messages. The default buffer size is 16k and the default debug level is 2. A debug level of 3 traces all calls in the module and a debug level of 0 suppresses all printk messages. Here is an example that overwrites the default values for buffer size and debug level:

  insmod ./fanout.ko buf_sz=8192 debuglvl=3

Fanout uses a kernel assigned major number so you need to look at /proc/devices to see what was assigned. The following lines create all ten of the possible instances of a fanout device.

  MAJOR=`grep fanout /proc/devices | awk '{print $1}'`
  mknod /dev/fanout c $MAJOR 0
  mknod /dev/fanout1 c $MAJOR 1
  mknod /dev/fanout2 c $MAJOR 2
  mknod /dev/fanout3 c $MAJOR 3
  mknod /dev/fanout4 c $MAJOR 4
  mknod /dev/fanout5 c $MAJOR 5
  mknod /dev/fanout6 c $MAJOR 6
  mknod /dev/fanout7 c $MAJOR 7
  mknod /dev/fanout8 c $MAJOR 8
  mknod /dev/fanout9 c $MAJOR 9

If all has gone well, the "Hello World" example given above should now work for you.

How /dev/fanout Works

This section is a high level design review of the fanout module. We explain the design and architecture, and relate specific lines of code in the module to the overall design.

The key to understanding how fanout works is to know how a little about how read() works. If you were to open a disk file and make five read() calls with each call reading a thousand bytes, you would expect the next read to give you the data starting with byte 5000. Internally, the operating system keeps a counter, called f_pos, that remembers where you are in the file. Once you've read the first 5000 bytes, you don't normally want to read them again, and since you aren't likely to ask for them again, fanout can forget them. The mechanism used to remember only the most recent data is a circular queue.

The fanout device uses the count variable to keep track of how many bytes have been written so far. At quiescence, the readers have all read count bytes (count and f_pos are equal), and the readers are now asking for data starting at *offset (which also equals count). Readers and fanout track bytes written

When a writer adds data to the queue, the count variable is incremented by the amount added. Each of the readers must now wake and read the bytes between *offset and count. After adding data to the queue, a writer wakes any sleeping readers with the call to wake_up_interruptible() in fanout_write().

Buffer overflow: One of the fundamental decisions to make in a design is what to do when a reader can not keep pace with the writers. In many designs you would apply flow control to the writers to slow them down to keep pace with the slowest reader. The fanout device, however, returns an error to the slow reader. Specifically, the reader gets an EPIPE error when it requests data that is no longer in the circular buffer (i.e. *offset < count - buf_sz, where buf_sz is the number of bytes in the circular buffer).

A reader does not immediately get an EPIPE after opening a fanout device that's been operating for awhile because in the file open routine, fanout_open(), we explicitly force the reader to be caught up with the writers. The line of code that does this is:

    filp->f_pos = fodp->count;

Code notes: It is said that programmers can read code and know what it does, but they can not read a variable and know what it means. So instead of reviewing the code, we are going to review the variables.

The fanout module supports up to NUM_FO_DEVS instances of a fanout device. NUM_FO_DEVS is currently set to ten. Each instance of a fanout device is described by the following data structure:

struct fo {
    char     *buf;           /* points to circular buffer */
    int       indx;          /* where to put next char recv'd */
    loff_t    count;         /* number chars received */
    wait_queue_head_t waitq; /* readers wait on this queue */
    struct semaphore wlock;  /* write lock to keep buf/indx sane */

Let's look at each of these variables in turn:

buf: The buf variable points to the start of the buf_sz number of bytes allocated for the circular queue. The memory is not allocated until the first open() on the device, and the memory is allocated using kmalloc(). Allocated memory is not freed until the module is unloaded.

indx: This variable gives the location of where to place the next byte in the circular queue. It is updated by fanout_write() as bytes are added to the queue. When indx gets to buf_sz, it wraps back to zero.

count: This variable is the total number of bytes written to the device. It is updated only by fanout_write() and a reader has data to read when count is not equal to *offset.

waitq: When a reader has no new data to read it blocks until new data is available. Specifically, the reading process sleeps in a call to wait_event_interruptible(). The writer's call to wake_up_interruptible() causes the readers to wake and continue execution with the lines of code immediately after the wait_event_interruptible().

wlock: While writers are writing to the circular queue, there is a short time during which the count and indx variables are not yet consistent with the data in the queue. During this window of inconsistency another writer might run and inadvertently corrupt the queue. The wlock mutex prevents this by locking out other writers while one writer is updating the queue.

One final note on the code is the use of the *private_data in the file structure. Fanout uses this variable to store a pointer to the struct fo appropriate to that file. The FanOut Device Pointer (fodp) is usually retrieved at the start of a routine with a line of code like this:

    struct fo *fodp = (struct fo *)filp->private_data;

Known bugs: While there may be several implementation bugs, the one possible design bug is that fanout assumes that the file offset counter never wraps. This should probably be fixed.

Security and Obsolescence

Security: In our use of /dev/fanout we let a web server read directly from it so that web clients can be updated when an event occurs. Giving the web server direct access to a device file is considered a security risk. You generally don't want your web server to follow symbolic links and you don't want the file system with the web server root directory to allow device nodes in it. (The file system should be mounted with the nodev option.) The fear is that if an attacker breaks Perl, PHP, or some other component of the web server, the attacker might be able be able to create /dev/hda1, /dev/mem, or some other critical device.

We will be using /dev/fanout in an appliance where, after boot, we can at least drop the system capability CAP_MKNOD when we drop the other system capabilities.

Obsolete already? It is widely anticipated that one of the next releases of the 2.6 kernel will include two new system calls, tee() and splice(). These calls will probably make obsolete the approach used to build fanout, and might make /dev/fanout entirely obsolete. This might be a good thing since, from a security point of view, it might be better to create a one-to-many multiplexer as a variant of a FIFO that can be attached to a nodev mounted file system. More information on tee() and splice() is available in an article at the Linux Weekly News (http://lwn.net/). The article number is 118750 and you can get to it directly here.


Bob is an electronics hobbyist and Linux programmer. He is one of the authors of "Linux Appliance Design" to be published by No Starch Press.

Copyright © 2006, Bob Smith. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 122 of Linux Gazette, January 2006

Stepper motor driver for your Linux Computer

By Sreejith N

An introduction intended for people with no prior device driver knowledge

This article is intended for those newbie Linux users who wish to use their Linux-box for some real work. I will also share some interesting experiments that I did with my AMD machine.


Learning new stuff is fun, but can be a bit frustrating. So, you want to write a device driver. The name itself is high-tech! You have some skills in the C programming language and want to explore the same. Also, you've written a few normal programs to run as processes in user space, and now you want to enter kernel space - where the real action takes place. Why Linux device drivers? The answer is,

Although it is possible to learn device driver coding by reading some books and PDFs written by the masters, this is a complicated and time-consuming approach. We will take the quick and easy approach, which is:

Let's make an easy start with some fundamentals.

Stepper motor basics

Stepper motors are special direct-current (DC) motors, typically used in applications like camera zoom drive and film feed, fax machines, printers, copying machines, paper feeders/sorters, disk drives and robotics.

A DC stepper motor translates current pulses into motor rotation. A typical unipolar (single voltage) motor contains four winding coils. Applying voltage to these coils forces the motor to advance one step. In normal operation, two winding coils are activated at the same time, causing the motor to move one step clockwise. If the sequence is applied in reverse order, the motor will run counterclockwise. The speed of rotation is controlled by the frequency of the pulses.

Basic Stepper Motor Conceptual Wiring Diagram

A typical full step rotation is 1.8 degrees, or 200 steps per rotation (360 degrees). By changing the time delay between successive steps, the speed of the motor can be regulated, and by counting the number of steps, the rotation angle can be controlled.

Stepper Motor Full Step Signal Timing Diagram

Bit Pattern for Full Step Mode
Green Blue Orange Red Hex Output Value
Step 0 1 0 1 0 A
Step 1 1 0 0 1 9
Step 2 0 1 0 1 5
Step 3 0 1 1 0 6

Hardware ideas

The circuit diagram for the drive is shown below.

Unipolar Stepper Motor Drive Circuit

The circuit consists of four TIP122 power transistors (T1, T2, T3 and T4), 220Ω resistors (R1, R2, R3 and R4), 3.3KΩ resistors (R5, R6, R7 and R8), 1N4148 freewheeling diodes (D1, D2, D3 and D4), and one LM7407 buffer chip (IC1). The 7407 buffer used here is a hex-type open-collector high-voltage buffer. The 3.3KΩ resistors are the pull-up resistors for the open-collector buffer. The input for this buffer comes from the parallel port. The output of the buffer is of higher current capacity than the parallel port output, which is necessary for triggering the transistor; it also isolates the circuit from the PC parallel port and hence provides extra protection against potentially dangerous feedback voltages that may occur if the circuit fails. The diode connected across the supply and the collector is used as a freewheeling diode and also to protect the transistor from the back EMF of the motor inductance. The motor used in my experiments (and documented here) was an STM 901 from Srijan Control Drives.

During normal operation, the output pattern from the PC drives the buffer, and corresponding transistors are switched on. This leads to the conduction of current through those coils of the stepper motor which are connected to the energized transistor. This makes the motor move forward one step. The next pulse will trigger a new combination of transistors, and hence a new set of coils, leading to the motor moving another step. The scheme of excitation that we have used here has already been shown above.

How do we interface the hardware with the Linux-box?

You can use either the parallel port or the serial port for this purpose. We will be using parallel port as a digital interface between PC and the hardware (stepper motor drive). The parallel port can be considered as a register, and the I/O operations can be done simply by writing bit patterns (numbers like 0xA, 10, '1010', etc.) to this register. The base address of parallel port is 0x378. The PC parallel port is a 25 pin D-shaped female connector in the back of the computer. It is normally used for connecting computer to printer, but many other types of hardware for that port are available.

Parallel Port Pinout

The original IBM PC's Parallel Printer Port had a total of 12 digital outputs and 5 digital inputs accessed via 3 consecutive 8-bit ports in the processor's I/O space.

  1. 8 output pins accessed via the DATA Port
  2. 5 input pins (one inverted) accessed via the STATUS Port
  3. 4 output pins (three inverted) accessed via the CONTROL Port
  4. The remaining 8 pins are grounded

To read a byte (8 bits) coming into a port, call inb (port); to output a byte, call outb (value, port) (please note the order of the parameters). The data outputs are provided by an 74LS374 totem-pole TTL integrated circuit, which can source (+) 2.6 mA and sink (-) 24 mA. The best option for preventing damage to the port is to use optocouplers; by doing so, the port is completely electrically isolated from external hardware devices.

What is a module?

Modules are pieces of code that can be loaded into and unloaded from a running kernel upon demand. They extend the functionality of the kernel without the need to reboot the system. Device drivers are a class of modules which allows the kernel to control hardware connected to the system. In this article, I have written a simple device driver to control a stepper motor drive; now it is time to log on to your console and start coding your very first module.

Let's have an unusual start! You have written so many "Hello, World" programs. So this time something different - "No gates, No windows, it's open". You will get these words printed on your console when you insert your first module.

/*mymodule.c - The simplest kernel module.*/
#include <linux/module.h> /* Needed by all modules */
#include <linux/kernel.h> /* Needed for KERN_ALERT */

int init_module(void)
        printk("<1>No gates, No windows, it's open\n");
        return 0;/* A non-0 return means init_module failed; module can't be loaded.*/

void cleanup_module(void)

The "start" function in a kernel module is init_module () which is called when the module is insmoded into the kernel, and the "end" (cleanup) function cleanup_module() is called just before it is rmmoded.

Compiling kernel modules

Kernel modules need to be compiled with certain GCC options to make them work. In addition, they also need to be compiled with certain symbols defined. This is because kernel header files need to behave differently, depending on whether we're compiling a kernel module or executable.

  1. -C: A kernel module is not an independent executable, but an object file which will be linked into the kernel during runtime using insmod.
  2. -O2: The kernel makes extensive use of inline functions, so modules must be compiled with the optimization flag turned on. Without optimization, some of the assembler macro calls will fail. This will cause loading the module to fail, since insmod won't find those functions in the kernel.
  3. -D_KERNEL_: Defining this symbol tells the header files that the code will be run in kernel mode, not as a user process.
  4. -W -Wall : Enable All Warnings.

Makefile (2.4.24 kernel)

As an example, let's take a look at the options we're going to use in compiling the "stepper" module that we'll see a little later in this article:

TARGET := stepper
WARN   := -W -Wall
INCLUDE:= /usr/src/linux-2.4/include
#INCLUDE:= -isystem /usr/src/`uname -r`/include
all : stepper.o
#${TARGET}.o : ${TARGET}.c
        rm -rf *.o

You can learn more about make utility by reading "man make".

Anatomy of device drivers

There is a lot of documentation on the Web, including PDFs and ebooks, on device drivers; also, you can download some useful guides from The Linux Documentation Project website. For the time being, just read these points carefully; later, you can move on to some detailed references.

Now let's examine our code.

#define MODULE

#include <linux/module.h>
#include <asm/uaccess.h>
#include <sys/io.h>
#include <linux/fs.h>

#define LPT_BASE 0x378
#define DEVICE_NAME "stepper"

static int Major,i,j,k;
static int Device_Open = 0;

//static int pattern[2][8][8] = {
//        {{0xA,0x9,0x5,0x6},{0xA,0x8,0x9,0x1,0x5,0x4,0x6,0x2}},
//        {{0x6,0x5,0x9,0xA},{0x2,0x6,0x4,0x5,0x1,0x9,0x8,0xA}}

static int pattern[2][8][8] = {

int step()
        if(k<8) {
//              if(pattern[i][j][k]==0) {
//                      k=0;
//                      printk("%d\n",pattern[i][j][k]);
//                      k++;
//              }
//              else {
//              }
        else  {
                printk("%d\n",pattern[i][j][k]); /*#####*/
                k++; /*#####*/
        return 0;

static int stepper_open(struct inode *inode,struct file *filp)
        static int counter = 0;
        if(Device_Open) return -EBUSY;
        printk("Opening in WR mode...\n");
        return 0;

static int stepper_release(struct inode *inode,struct file *filp)
        Device_Open --;
        return 0;
static int stepper_write(struct file *file, const char *buffer, size_t len,
loff_t *offset)
        char *data;
        char cmd;
        switch (cmd=data) {
                case 'H':
                        printk("Reffer README file\n");
                case 'h':
                        printk("Half-Step mode initialized\n");
                case 'f':
                        printk("Full-Step mode initialized\n");
                case 'F':
                case 'R':
//              default:
//                      printk("Give 'H' for Help\n");
//                      break;
        return 1;

static struct file_operations fops={

int init_module(void)
        Major = register_chrdev(0, DEVICE_NAME, &fops);
        if (Major < 0) {
                printk("<1>Registering the character device failed with %d
                return Major;
        printk("<1>Registered, got Major no= %d\n",Major);
        return 0;

void cleanup_module(void)
Follow this link to download the code.

Driver initialization

  1. The init_module() function is called on the driver's initialization
  2. The cleanup_module () function is called when the driver is removed from the system
  3. The init function will register hooks that will get the driver's code called when the appropriate event happens
  4. There are various hooks that can be registered: file operations, PCI operations, USB operations, network operations. Ours is a file operation.
  5. The driver registers a character device tied to a given major number and a user can create access points corresponding to this major number.The following command will do it for you:
mknod /dev/stepper c 254 0

How stuff works

A user space program can write commands to the device file to rotate the stepper motor through a desired angle at desired speed. The speed of rotation depends upon the delay given in the user program.

The built in commands for controlling the stepper motor is given below.

File operations

The driver makes use of the following device file operations:

  1. open for allocating resources
  2. release for releasing resources
  3. write the required pattern to the parallel port.
  4. there is no reading in our program, but if you want, you can read the current pattern at the parallel port.

If you write 'F' once to "/dev/stepper", the motor will rotate through its minimum step-angle. If you keep on writing 'F' to "/dev/stepper", it will rotate continuously. The "write" system call will do this for you.

#include "header.h"

main ()
        char t,buf[6] = {'h','f','F','R','H','q'};
        int fd,rt,i,j;
        size_t count;
        printf("Select Mode \n(1) [Half-step clockwise]\n(2) [Half-step
anti-clockwise]\n(3) [Full-step clockwise]\n(4) [Full-step anti-clockwise]  :
        if(t=='1') {i=0; j=2;}
        else if(t=='2') {i=0; j=3;}
        else if(t=='3') {i=1; j=2;}
        else {i=1; j=3;}
        for(i=0;i<1000;i++) {
                usleep (100000);

Also, if you are familiar with shell scripting, you can do the same by writing a simple shell script. Now you can start talking to the device /dev/stepper. It will be really interesting if you talk in Linux's language - I mean a shell script. Just use simple echo commands as given below:

echo H > /dev/stepper

Do you think that Morpheus is talking to you? Your kernel is replying to your commands. How's that! Now, you too can feel like you are The One.


I hope I have given you some basics of device driver coding and perhaps a "small step toward Robotics". Here is a detailed schematic of a stepper-controlled robotic arm; feel free to try it out. One can connect three stepper motors simultaneously to the PC parallel port and can achieve step-wise mobility; this allows anyone to start thinking of complex innovative mobility with multi-threaded programming in Linux. You can also add C functions to our module to enhance its functionality... the possibilities are endless!


I am a Linux enthusiast living in India. I enjoy the freedom and power that Linux offers. I must thank my mentor Mr. Pramode C. E. for introducing me to the wonderful world of Linux.

I completed my B-Tech in Electrical and Electronics Engineering from Govt. Engineering College, Thrissur (Kerala, India) (2001 - 2005). Presently I am working in inDSP Audio Technologies Pvt. Ltd, Trivandrum, India as an Embedded Systems Engineer.

I spend my free time reading books on Linux and exploring the same. My other areas of interest include device drivers, embedded systems, robotics and process control.

Copyright © 2006, Sreejith N. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 122 of Linux Gazette, January 2006


By Shane Collinge

These images are scaled down to minimize horizontal scrolling.

Click here to see the full-sized image


Click here to see the full-sized image


Click here to see the full-sized image


All HelpDex cartoons are at Shane's web site, www.shanecollinge.com.

[BIO] Part computer programmer, part cartoonist, part Mars Bar. At night, he runs around in a pair of colorful tights fighting criminals. During the day... well, he just runs around. He eats when he's hungry and sleeps when he's sleepy.

Copyright © 2006, Shane Collinge. Released under the Open Publication license unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 122 of Linux Gazette, January 2006

The Linux Launderette


(?)Lots of OT stuff
(?)Fwd: Xteddy. :)
(?)Microsoft and open standards
(?)Happy $HOLIDAY
(?)Email I did not send
(?)Dilbert blog: Winning by Knockout
(?)Charity sweepstakes spam
(?)Inuit words for snow
(?)You just gotta laugh!!
(?)Mozilla search plugin for LG
(?)Crap music
(?)Japanese help

(?) Lots of OT stuff

Wed, 28 Dec 2005

From Jimmy O'Regan

I didn't get the launderette in on time last month, so last month's is available.
Over the past few months there have been quite a few recurring topics. Two of them have been the law and Web 2.0
(Also, in the latter thread, I mention some XSLT for photo annotations. Here's an example of that)

(?) Fwd: Xteddy. :)

Mon, 28 Nov 2005

From Stefan Gustavson

(!) [Thomas] I am liking this. :)

Thanks for the XTeddy fan mail. It's always nice to hear from people who like him. I don't get more than a few messages per year about him these days, but it's nice to see that he's still out there.

The real life XTeddy is a present from my wife, she got it for me as a stress relief in 1994 when I was a PhD student, and I must say it still works. He still sits on my desk, or rather on a shelf just above my desk, in clear view and ready to comfort me when things get rough. (Although his mere presence actually seems to stop things from getting too rough.)

Like most teddy bears, he has a real life name, and it happen to be "Lufs", which is a cutesy Swedish bear name whose point gets more or less lost in translation, but I'll try anyway:

The name "Lufs" means nothing in particular, but the kind of slow trotting and wobbling walk that real live bears tend to do when they are not in any hurry is referred to as "lufsa" (a verb) in Swedish, and when we were kids, both my wife and I had read a story by the Swedish author G=F6sta Knutsson about a little bear cub named "Nalle Lufs", ("Nalle" is an old Swedish word for "bear", often used as a synonym for "teddy bear".)

I can't remember whether it was me of my wife who named him. It kind of grew on him more than being a planned naming. I'm not even sure if we were consciously aware of that childhood story at the time, but we both recalled reading it when we saw it a few years ago.

The XTeddy program was my first and so far only attempt at writing a native program for X Windows. It makes very simple use of the Shape extension and a pixmap background, and the original version was about one page of code. It has since been extended with better image loading functions, but the basic window functionality is still just a shaped window with a background pixmap, and a simple, low level mouse event handler to move it around.

It was also a counterstatement to "xeyes", the only other shaped window program that was around in 1994. I thought it was a bad and ugly demo for a good extension. Xeyes is kind of cute the first time you try it, but it is kind of intrusive (someone or something is watching your every move with the mouse) and therefore has very little staying power with most users.

XTeddy is rather the opposite to xeyes: Nonintrusive, motionless unless you pick him up, doesn't really care what you are doing and watches you, not your work or lack thereof, in patient silence. He's basically useless, but cute, and always available.

That was probably more background information than you ever wanted, but you did ask for it. :)

(!) [Thomas] xteddy has been personified into a bot that gives people hugs.

(?) He's very good at that. Good thinking! :)

(?) Microsoft and open standards

Tue, 06 Dec 2005

From Jimmy O'Regan

"Microsoft has stated on numerous occasions that they believe in and support open standards. But from my experience, they do this not by joining existing open standards efforts, but instead by creating entirely new, parallel (and arguably redundant) "open standards" efforts around their own technologies."


(?) Happy $HOLIDAY

Sun, 25 Dec 2005

From tag

Up to you whether I'm late, on time, or early for your personal favorite.

(!) [Ben] I've been running behind on everything - Ulysses is in the final stages of the repairs, and we're spending all our time sanding, grinding, and painting (and learning a hell of a lot, particularly about that last; there's a whole body of science behind it, particularly when you ask as much of it as we do.) Anyway - happy holidays to everyone, whichever version you celebrate; may your upcoming year be filled with love, joy, and happiness!
(!) [Jimmy] And may you have fewer hurricanes!

(?) And my apologies to Jimmy for the #tag-chat teddybear being asleep on the job. Apparently he knows how to serve up the brews right proper, but only serves them correctly on one network. D'oh!

(?) Actually I sit corrected; poor little plushybrains watched his pearl roll under the couch. I had to unload and restart his script, and then it worked fine in both places.

(!) [Jimmy]
[18:55] <jimregan> Um... Happy cheesy /topic messages to you all too
(!) [Jimmy] For the record:
* Topic is 'The Answer Gang, happy linuxgazette.net to all penguins
everywhere | +cgianakop> I just wanted to stop in for a few and wish the
gang merry christmas | merry solstice, happy hanukah, and a cheerful xmas
too | xteddy serves eggnog and fine brews | happy gnu year'
(!) [Jimmy]
[19:01] <jimregan> xteddy: beer Heineken
[19:01] <xteddy> :)
[19:01] <jimregan> xteddy: beer Heineken please
[19:01] <xteddy> :)
[19:01] <jimregan> xteddy: beer please
[19:02] <xteddy> :)
[19:02] <jimregan> xteddy: hi
[19:02] <xteddy> :)
[19:02] <jimregan> xteddy: eggnog

(?) Thanks to my attempts to merge some things, at the moment it's "give me" or gimme.

xteddy, give me a beer
xteddy: gimme Heineken
xteddy give me Heineken please

...although the result on this one will look a little silly he doesn't know "hi" so all you get is a smily

xteddy, give me eggnog
(!) [Jimmy]
[19:02] <xteddy> :)
[19:02] <jimregan> xteddy :-P
[19:02] <xteddy> :)
[19:02] <jimregan> xteddy: Heineken
[19:02] <xteddy> :)
[19:03] <jimregan> xteddy: poprosze wodke
[19:03] <xteddy> :)
(!) [Sluggo] If I threw xteddy out the window, would I incur the Wrath of Thomas? Would my Linux questions go unanswered forevermore?
(!) [Jimmy] Hmm. Wasn't that the little known Egyptian Plague 6.5? Boils in the shape of Xteddy?

(?) is this the verb for "give me" ? I could add it.

(!) [Jimmy] No, well, kind of... In a bar, you'd say 'proszę <whatever>'. It's probably better for xteddy to not understand Polish, because of the amount of grammar you need to know just to ask for something :)
Well, maybe not. A sample session'd look like this:
<jimregan> xteddy: prosze wodke (vodka please)
<xteddy> jimregan: masz wodke (your vodka)
<xteddy> jimregan: wodki nie ma (out of vodka)
<jimregan> xteddy: prosze piwo (beer please)
<xteddy> jimregan: masz piwo (your beer)
<xteddy> jimregan: piwa nie ma (out of beer)
(!) [Ben] It means "I would like", if I understand it correctly; not too far off a similar Russian word, although the usage is somewhat different.
(!) [Jimmy] Kinda... it means "I request". It's the perfective version :)
(!) [Jimmy] For those who are not familiar with perfective verbs: Wikipedia (as ever) has a good article.

(?) There's precedent. He already does hug in a few foreign languages for some specific people. As I continue to merge his code bits he'll be a multilingual teddybear.

(!) [Jimmy]
<xteddy> przytulę <whoever>
* xteddy przytuli <whoever>

(?) I'm not doing it i18n style yet though.

(!) [Jimmy]
<xteddy> przytule <whoever>
Or you could use tule/tuli (1st/3rd person of 'cuddle')
(!) [Jimmy]
[19:03] <jimregan> Hmm. Xteddy nie mowi po polsku
[19:03] * xteddy bounces

(?) he saw his proper name... the trigger.pl script wasn't busted, making the bear look hard of hearing rather than simply not there.

(!) [Jimmy] I thought it was something like that :)
(!) [Jimmy]
[19:03] <jimregan> xteddy: czy mowisz po polsku?
[19:03] <xteddy> :)

(?) What's this mean?

(!) [Jimmy] "Do you speak Polish?"
(!) [Jimmy] And there I was thinking xteddy was trying to tell me something :)

(?) Like what, have a guiness instead ? :P

(!) [Jimmy] Nah, like "I think you've had enough, and I think you should give me your keys. No, I don't care that they're house keys -- you're in no fit state to operate any form of door".

(?) Anyways, feature request to have the recognized goodies work without "gimme" is already in the works... I wanted to merge the old features first, but the gimme was too attractive an idea.

Entirely aside of teddybears, it's a nice warm feeling to see progress, so I think it was worth it. For giggles, people can try a few types of snacks, drinks, and baked goods.

Chris G. came in while I was still coding up some of this stuff, and followed me into #fvwm to see him in a little more action (the regulars there are pretty found of our bear). He was delighted.

(?) Which is a damn shame, because I've almost got him handling guiness with proper style too.

(!) [Jimmy] <jimregan> :)

(?) I'm off to enjoy nog and partying now...

(!) [Jimmy] Well... Wesołych Świąt, Maligayang Pasko, Nollaig Shona Daoibh, or just plain Merry Christmas.

(?) sudden burst of on-topic: something's eating those characters, and I'm mildly sure it's not mutt, I know I've see diacriticals and odd chars in mail from some people now and then. Did anyone among us get that first couple of words in correct form, and if so, how do you think it did it?

(!) [Ben] Various 10646 fonts suck in arbitrary ways; you're probably missing the glyphs for Polish in whatever you're using. All the stuff in Jimmy's email displayed fine for me in Mutt. Here are my relevant bits:
r!egrep '^xterm.*(utf|font)' ~/.Xresources
xterm*font: -misc-fixed-medium-r-normal--18-120-100-100-c-90-iso10646-1
r!echo $LANG
I used to have to mess about with LANG and LC_ALL and a whole bunch of goodies to get Mutt to act right...
(!) [Breen] This may be a project for me in the new year - I'm currently reading most mail on shell on a FreeBSD box, using mutt displayed on various terminals on FC3 and Mac OS X. Getting the correct invocations to make this all work hasn't been near the top of the list, but this may be the year...
(!) [Ben] You're in luck: I still have the wrapper script I used to use to launch Mutt to read my Russian mail, which contains the previous invocations (commented out) to do various language-related tweaking. I still use it, since I like saving attachments to /tmp by default and prefer opening Mutt in "mailbox" view ('-y').
# Created by Ben Okopnik on Tue Sep 11 14:41:52 UTC 2001

cd /tmp
# LANG=ru_RU rxvt +sb -geometry 100x35+-5+-5 -fn 10x20-koi -n RMutt -e mutt -y
# LC_CTYPE=ru_RU xterm -fn 10x20-koi -e mutt -e "set charset=koi8-r" -y
# LANG=en_US.UTF-8 xterm -fn terminus-20 -T RMutt -e mutt -y
# LANG=en_US.UTF-8 xterm -fn -misc-fixed-medium-r-normal--18-120-100-100-c-90-iso10646-1 -T RMutt -e mutt -y
xterm -T RMutt -e mutt -y
(!) [Breen] Happy New Year to all!
(!) [Ben] now, using the above font and $LANG variable for all of X takes care of it just fine. And ':dig' in Vim produces a very satisfying result. :) About the only language that I had trouble displaying was Georgian - 'gnu-unifont' handled it, but 'gnu-unifont' sucks when you try to display anything with line-drawing characters (i.e., Midnight Commander). Oh well; 99.9% functionality is miles better than none.

(?) Email I did not send

Fri, 16 Dec 2005

From Vladytoe

See attached aoler.html

(!) [Rick] How... nostalgic. That seems to have been an instance of an actual confused AOL user, writing manually from AOL, rather than a spam forgery or malware mail pretending to be from AOL. (All of the headers verify this, including SPF, which AOL implements properly.)
What's completely unclear from the information given is why Vladytoe@aol.com is writing, since her headers show only intra-AOL mail previously.

(?) Dilbert blog: Winning by Knockout

Sun, 25 Dec 2005

From JimRegan

I always consider myself a winner by knockout whenever someone distorts my point to something ridiculous and then argues against the ridiculous thing. That seems like an acknowledgement that my real point is unassailable. Why else would someone need to invent a whole new point to argue against?
The Dilbert Blog: Winning by Knockout

(?) Charity sweepstakes spam

Sat, 17 Dec 2005

From Mike Orr

This is to inform you that, base on our end of year promotion of The
Programmer Charity Foundation carried out on the 15th
,December,'05,the above ticket No. fixed to your email ID surfaced in
the third ballot stake.

You have therefore emerged as a recipient of the third award of two
hundred and fifty thousand american dollars in this annual end of year

I wonder if they pasted the amount out of a 419 email?

The source of fund is majorly from well meaning companies and
humanitarian spirited industries selected from european community.

Be informed that in line with the sponsors' regulation,you are to
contribute a percentage amount( ten percent minimum) to a recognized
non-governmental charity project in your environ,after collection of
your prize

Any charity or a certain one? What does "my environ" mean?

Application and inquiries should be forwarded to Mrs Mabel Konnindam
via ; mabel@programmer.net.

Paragraph 2 says I've already won, so why do I need to send an application?

Accept my heart felt congratulations.

Ms Daisey Gregson
PS:Kindly state your Ticket No in your expected application.

The Programmer Charity Foundation,
Schalktorrenstraat  673b,
Royal Dutch Federation.

(?) Inuit words for snow

Fri, 18 Nov 2005

From Jimmy O'Regan

We've all heard that the inuit have 20 something words for snow, right? Well, here's a page listing them: http://www.geocities.com/Athens/Forum/7953/eskimo.html

I found this from "The Czechs Have a Word For It": http://www.bohemica.com/?m=catalog&s=258&a=138 having looked up "Nedovtipa" ("one who finds it difficult to take a hint"), after seeing a mention in "The Meaning of Tingo" -- well worth a read.

Some great words are:

  • "neko-neko", Indonesian for "one who has a creative idea that only makes things worse"
  • "bakkushan", Japanese for "a woman who appears pretty when seen from behind but not from the front"
  • "scheissenbedauern", German for "the disappointment one feels when something turns out not nearly as bad as one had hoped" (literally "shit regret")
  • "tsuji-geri", Japanese for "trying out a new sword on a passer-by"

(It also mentions that the Albanians have 27 words for both eyebrows and moustaches).

(!) [Heather] [neko-neko] Might be an amusing name for a wiki about dumb things you shouldn't try at home.
As a multilingual amusement, neko is japanese for kittycat, and many kittycats have occasioned exactly this kind of brilliant idea.

(?) There's a website: http://themeaningoftingo.com It has a list of links to similar sites: http://www.goodguys.co.uk/tingo/links.htm

Another word: "Backpfeifengesicht", German for "a face that cries out for a fist".

It turns out that "scheissenbedauern"'s not a real German word:


It turns out [the Alb?nian words] are just adjectives that are added to the words for moustache and eyebrows, not separate words.

My favourite Polish word is 'kobietka', the diminutive of the word for woman, that roughly translates as "little girl pretending to be a woman". We have an equivalent word for men in Hiberno-English: "maneen".

(!) [Ben] The Russian equivalent for the latter is "muzhichok";

(?) I know there's a Polish equivalent, but I can't find it. The only diminutive I've heard applied to men is 'dziewcinka', and that means 'little girl'.

(!) [Ben] oddly enough - since I would have expected two closely-related Slavic languages to share a similar concept - I can't think of a parallel to 'kobietka'. There's "suchka" ("little bitch", a fond diminutive sorta-insult), but that's not the same thing.

(?) That's 'suczka' in Polish, so no :)

(!) [Ben] Yep, I knew that one.

(?) I assume there's something like 'żoneczka' ('little wife')

(!) [Ben] No, the diminutive for 'wife' doesn't have any extra connotations besides the obvious ones.

(?) Oh, I was just translating literally -- no extra connotations here.

(?) You just gotta laugh!!

Fri, 25 Nov 2005

From Pete Savage

Hi all, I'm sorry I just had to post these here, my father-in-law almost paid for a Navman ICM 550 off of eBay, but being a shrewd man, he did some feedback chcking and immediately noticed something was wrong. I looked at these and oh......my sides doth split. I don't know if they have constructed a mechanism for creating comments from a load of random words, but the poor translation says it all.

http://feedback.ebay.co.uk/ws/eBayISAPI.dll?ViewFeedback&userid=yd6s7 http://feedback.ebay.co.uk/ws/eBayISAPI.dll?ViewFeedback&userid=corinalucia

Visit them soon, for you never know when they will be taken off eBay.

Here's a sneak preview

Positive feedback rating - I will chase the friend whom your item introduce to me of.

Hope you all get a laugh from this

(!) [PeteJ] Oh my sides are they splitting!
May I fondle your nipples?
(!) [Sluggo] "You are a seller of honesty." ???
That would make a great cartoon, that one.

(?) I think this is one of my favourites

Useless buyer,is an ebay humiliation not at live under go to, you choose suicide

I wonder exactly what the meaning was supposed to be, if there was a meaning at all.

(?) Mozilla search plugin for LG

Mon, 21 Nov 2005

From Jimmy O'Regan

Install with: javascript:window.sidebar.addSearchEngine("http://linuxgazette.net/122/misc/nottag/lg.src", "http://linuxgazette.net/122/misc/nottag/lg.png", 'Linux Gazette', 'Webzine');

(Be warned that it may not work, and may even royally screw up your browser)

See attached lg.src

(?) Oh yeah: works for me. (Only tested in Flock, though)

(!) [Thomas] So much work, for such a large browser. :) For those of us that use Elinks as their browser, then here's something similar, using "Smart Prefixes". Essentially:
Options Manager -> Protocols -> URI rewriting -> Smart Prefixes
... from there, one can "Add" a new prefix -- I called mine 'lg'. When that's done, scroll down to it, and "Edit" it, adding the following to it:
... the "%s" variable above expands to whatever search term was entered in. Then click on "Save", and voila. Now one can search LG, by either pressing 'g' or 't', and typing:
lg: my search term
More information can be found on my own Elinks site here:

(?) So much work, for such a small browser :)

(!) [Thomas] Hehehe. :)

(?) Mozilla &c. can do the same thing: bookmark the above URL, and add 'lg' as the keyword. Searching is similar, but without the colon.

(!) [Thomas] Well, the colon is optional -- the delimiter is either the first space after the keyword ("prefix" as I would call it), or a colon.

(?) The point of the file I posted is that it adds the search to the little search bar at the top of Firefox.

(!) [Thomas] Yeah - it's very cute. :) I suppose it's just preference at the end of it all.

(?) Nah, it's laziness: I can't be arsed remembering the keywords for different searches :)

(?) Crap music

Tue, 29 Nov 2005

From roadtripper8 at 43 Things

(!) [Jimmy] This started here where the theme music from "The Heights" was mentioned
(!) [Jimmy] To where you are.
It's like trying to catch a falling star.
*Don't* blame me -- you started it!

That has to be the crappiest song ever written. And yet we still remember it!

(!) [Jimmy] No, I can think of worse
"Barney is a dinosaur..."
"I want to be the very best, like noone ever was"
All TV themes. Eerie.
No, wait... Celine Dion's back catalogue!
Funnily enough, I left my MP3 player at home last Friday, and that song popped into my head for the first time since the show was on. Your comment... well, it reopened a fresh wound. Damn you!

(?) Make it stop!

I spoke too soon. You have managed to find songs that really are crap.

Hopefully your revenge won't sneak up on me over the next few days. 'Cause I may have to throw myself in front of a truck if Celine Dion starts playing in my mental jukebox.

(!) [Jimmy] It's not hard to find crap music -- just look at the charts. (Quote: "There's more evil in the charts than an Al Qaeda suggestion box" -- Bill Bailey)
Not quite unrelated: my sister has been a bit upset with me since MTV played all of Kelly Clarkson's videos and I told her they sounded like Diet Coke ads. And she had to admit they did. (At least I didn't have to admit I thought one of them was catchy).
Oh, and as for poor old Celine: I shared a house with one of my friends for a year in college--without checking his CD collection first. 9 CDs, all Celine Dion.

(?) I roomed once with a Backstreet Boys fanatic. I when I say fanatic, I mean she actually put up a poster of them...*IN OUR BATHROOM!*


(!) [Jimmy] You have bested me
I can normally think of a smart ass comment for every situation, but... overload!

(?) Consider it payback

I am the one who had the Barney song going through my head for a day!

(?) Japanese help

Fri, 18 Nov 2005

From Pete Savage

Hi all,
A lot of you seem to be well up on foreign languages.
I need to find out how to "write"
"Lisa, I love you." In Japanese.
I think I'm ok with the Lisa part, needs to be written in katakana as Risa.
Once again Any help greatly appreciated.
(!) [Kat] Well...Pete, why do you want to write this in Japanese? I can help you say this in Mandarin, but the Japanese just don't say this. It sounds horribly stilted and overblown, even though there is a literal translation. Even the more commonly used declaration sounds well, kinda dorky.
So with those caveats,
"Lisa, anata aishiteru" is probably as good as it's going to get. (Lisa, you(familiar) loving) (If you add in the "I" , it sounds like the emphasis is on I love you, rather than "you are loved by me".)
"Suki da yo" is what a guy tells his girl, but IMHO it has connotations of "yeah yeah, I'll TELL you I like you because you asked, but it should be obvious")
(!) [Jay] Oh, so Japanese guys are just like American guys. Got it.

(?) Thank you for the reply.

It's for my wife for christmas, she is very fond of the Japanese culture, in particular Manga and Anime. I wanted to write her almost like a scroll that I could frame for her for christmas

(!) [Rick] Well, good for you.
I personally had to sit down, a bit over five years ago, and figure out how to write?"Deirdre, will you marry me" in Esperanto.
(!) [Thomas] Wow, it took you all this time? :)
(!) [Rick] She, in turn, set herself the task of figuring out the equally daunting challenge of how to say "Yes" in Irish.
(!) [Jimmy] Ooh! I just sprayed cereal all over the place! Did she manage to find something shorter that "I will marry you", or did she go with "is d=F3cha..." ("I suppose...") :)
(!) [Rick] She did indeed go with "I will marry you."
(!) [Sluggo] I suppose this "I suppose" sounds better than it does in English? ("Why not, I have nothing better to do today?" "Nobody better has come along yet, so what the hell?")
(!) [Rick] That does remind me, as it happens, of Deirdre's favourite Irish joke:
John and Mary had known each other, and gone out to dinners and dances, for well on twenty years. Mary, being rather shy, eventually gets around to raising the subject, reminding him of this and adding:
"John, don't you think we should be marrying, then?"
"Aye, but who will have us now?"
(!) [Jimmy] Um... OK, when you were in Ireland, you must have heard the phrase, "sure why not", or "sure we may as well" (where 'sure' is pronounced 'sher' and not emphasised)... it'd sound like that.
(!) [Sluggo] I don't remember that but I've heard "sure" alone in Scotland (and in the movie Local Hero ) . Quiet and monotone. I'm used to a strong "sure" meaning "I really like the idea",
(!) [Jimmy] Sure :)
(!) [Sluggo] a short but still forceful sure being concessionary ("OK, that's fine"), and a long up-down-flat "suurrrre" being ironic ("Sure you're going to the mall" [but I don't believe it]). A quiet, flat "sure" to me sounds very concessionary, as in "I don't want to do it but I'll go along with your wishes." I had to tell myself, "That's not what they mean," and trust the words rather than the intonation.
(!) [Jimmy] Um... I'm not sure (sure sure, not sure sure :), but I think the Scottish use it the same way as we do. It doesn't really mean anything, it's kind of an introduction to a sentence -- in fact, most of the time it means you're unsure. Most Irish-type uses could be replaced with "well" (or even "uh..."): "sure, you can take a later train if you miss this one".
(!) [Sluggo] That's a different context. I'm talking about a one-word response to a proposal or a yes/no question. In Local Hero , when the Texas oil negotiator gets off the plane in Aberdeen and meets his assistant Olson, he asks about logistics like "Do you have a car?" Olson replies, "Sure." The negotiator is surprised that Olson just says it like a boring fact ("Apples grow on trees") rather than anticipating that the question imples, "Let's go." So half-exasperatedly he says, "Well, let's go then." Probably Olson did not mean it that way, but that's how it sounds to yankee ears, extremely detached, like the speaker really doesn't care or something. It's hard to describe because we just don't say "sure" that way, so we interpret it like other phrases that are said that way.
(!) [Jimmy] (IIRC, there's a region in Poland (Poznan, possibly) where they do something similar ("??ucho", I think)).
(!) [Sluggo] What struck me about Ireland was how so many accents sounded indistinguishable from American. Only if you listen hard could you hear faint traces of "foreign" intonation. At least a quarter of the people I met were like that, if not a third.
(!) [Jimmy] Heh. You obviously weren't in Cork, North Dublin, Cavan, or rural Tipperary, to name but a few of the areas with indecipherable accents. I'm from rural Tipperary, but I rarely stand a chance.
Partly it's because there's a Hiberno-English tendency to speak short sentences as if they were single words: "bulnfrish" or "blinfrit" are two renderings I've heard of "bullin' for it"[1] (i.e., horny).
(!) [Sluggo] Like Mumbles in Dick Tracy.
(!) [Jimmy] [1] Thankfully, we managed to contain that phrase before it could infect other areas, though there is still a high risk of breakout.
(!) [Brian] I would suspect that the easiest route would be to head over to Ireland, get out in the countryside a bit, stop in pub and ask the first person you see if they'd like you to buy them a pint. Virtually any answer you get there should be fairly equivalent to "Yes!".
.brian (the empirical linguist)
(!) [Sluggo] Just be sure not to say "I will accept a beer" when you meant "I will marry you".
(!) [Jimmy] Well, you know, think of Vegas weddings and it's not much of a stretch...
So, the more I read of this "Meaning of Tingo" book, and the accompanying blog, the less I doubt the veracity of anything I read in it,
(!) [Ben] Do you mean "less" or "more"? Those double negatives are tricky. :) As to "veracity", it doesn't apply to books; veracity is a habit, not an absolute measurement. "Accuracy", maybe?
(!) [Jimmy] sigh Yeah. I'd just woken up, and my brain still hasn't started working yet.
(!) [Jimmy] especially as the only Irish word so far is: "nabocklish: don't meddle with it", which is... well, that may be a Hiberno-English word in some part of the country, but "ná bac leis" means "don't pay attention to *him*".
(!) [Ben] Ya gotta figure: in a book like that, effect is very important, so bending the truth a bit to achieve it, well, it's not really that big of a stretch...
(!) [Jimmy] Plus, it has to be taken into account that linguists are at the mercy of the sense of humour of the person they are questioning; especially difficult on the internet, where you can't see if they're having difficulty keeping a straight face (one of the Philipinos at work told me that "putang ina mo" means "hello", for example. If you're really good friends with the person you're talking to, maybe...)
(!) [Jimmy] So, Kat, can you confirm that "ariga-meiwaku" means "an act someone does for you that you didn't want to have them do and tried to avoid having them do, but they went ahead anyway, determined to do you a favour, and then things went wrong and caused you a lot of trouble, yet in the end social conventions required you to express gratitude", because I can think of several times where that would have summed things up :)
(!) [Kat] Other than the fact that I know this word as "arigato meiwaku", yes, that's about the gist of it. "meiwaku" means bother/annoyance/imposition.
Of course the etymology of "thank you" in Japanese is peculiarly Japanese in concept as well. Quoting from someone quoting the sci.lang.japan faq:
The word derivations are different too. Arigatou is related to arigatai .kind, welcome. and arigatagaru .be thankful., which conveys a grateful feeling for something that was done in favor of the one using the word. The first kanji is aru (. Unicode 6709) .exist, happen. and the second one, gatou < katai (. Unicode 978F) means .hard, difficult..

(?) What would be most appropriate in this instance

(!) [Kat] Eek! That's really sweet, but in a Japanese context it's even more bizarre. Declarations of love like that are meant to be private in Japan, so something that would be framed and on a wall... Aiee.
(!) [Jimmy] At least he's not going for a tattoo :)
(!) [Kat] I'm not going there. I can't afford the bus ticket back.
(!) [Jimmy] Heh. I've been planning to geta tattoo of my son's name in Ogham. After 8 years of research, I'm reasonably confident that what I've chosen is his name and not, say, "Price reduced for a limited time only" or some such.
(!) [Kat] The trick now is to make sure it says "name", rather than "nayyun" or some such close but not quite thing.
I'm still trying to figure out how someone got "Opoka" out of "Okopnik".
(!) [Sluggo] My Russian roommate Was Valeriy Pronin (vah-LEHR-ee PRO-neen) but the border guards anglicized it to Valerie Pronine. So not only was he stuck with a girl's name but people pronounced the last name "pro-nine".
Then he married an Irish-American girl and changed his name to Michael Sean Kennedy. That really throws people.
I went to Russia and his family met me at the airport. I was expecting the Pronins but found the Dmitrochenkos. Turns out his nickname was his given name. He and several buddies had changed their name to Valeri Pronin after a buddy by that name died. I was irritated that he hadn't told me his family's name was different than his, because if his brother hadn't recognized me at the airport I'd be looking around for Pronins and never have found them. Which was scary because it was my first time abroad and Russia is not a very familiar place. ("What if I get to my hostel at midnight and it's closed?")
Another friend was named Chris Norman. I went to visit his cousin in Russia. Since it was his cousin I expected a different name. They met me at a metro station in Moscow so it was easy to find them. (I still don't like taking Russian suburban trains alone. So easy to take the wrong train, and good luck recognizing the station when they announce it.) Later in St Pete I saw posters for Chris Norman, a musician, so I don't know if he did something funny with his name too.
(!) [Ben] Heh. That would be a local dockmaster (from the UK, originally), with an interesting policy on spelling:
(me) -- "Just backspace until you get to 'O'."
(him) -- "Naow, naow, mite. Oi *never* goes backwards, only forwards."
         "There, 'at should do it!"
(!) [Kat] I'd swear it was a hand-written receipt, though.
That's okay, at least it wasn't like having one's drycleaning filed under "rat".
(!) [Kat] Would she go for a Chinese hanzi (i.e. looks like Japanese kanji) thing instead?
What you'd want in that case is "Wo ai ni", which is just fine.

(?) Yes that'd be fine, what is the literal translation?

(!) [Kat] I love you, in Mandarin Chinese.
I should state rather strongly at this moment that I think you should find a Chinese calligrapher to do this, as most people who are not learned in the writing of Chinese characters make an embarrassing botch of it. You can try googling "kanji tattoo" to see what most native speakers think of the currently trendy usage of kanji.
(!) [Rick] Using a language in a culturally inappropriate way can actually be an art form, y'know. ;->
(!) [Jay] "Somehow, I knew who said this, without even looking at the attribution."
(!) [Sluggo] Exactly, and haikus have developed an independent life in English that doesn't always fit with the Japanese tradition.
(!) [Rick] The interdependencies of language and mind are always fascinating, as I started to realise when I learned French and realised that there were particular modes of thinking that come naturally when "thinking in French" but not in English, and vice-versa. (I'm straining to remember examples, and, unfortunately, failing.)
(!) [Ben] I can't swear a tenth as well in English as I can in Russian. It doesn't roll off the tongue, it doesn't sound natural, and... it's not polite. %->

(?) I love the russian langauge, don;t know squat, just like the way it sounds

(?) I wonder if there are other phrases that may write better if you know what I mean? Sound more traditional, even though you said declarations like this are private, they must be "said" ? I know not. I understand little.

(!) [Kat] You could do "Lisa daisuki" (Lisa (is) (my) favorite), which has a certain "adolescent carved on a tree" charm to it.
I suppose the traditional thing would have been to compose a cryptic impromptu haiku.
(!) [Rick] There you go! Pete, if I were you, I'd fire up my mad haiku (or at least senryu) skillz.
(!) [Heather] Speaking as a girl here a haiku sounds sweet ;)

(?) I have a haiku

Beauty stealing breath,
Grace compared to angels high,
Always in my heart.
(!) [Sluggo] That's good. She'll know what it means. it's like the one we have about Tux. Hints at it without explicitly saying it.
Bottom of the world
Fat, dark and handsome
Herring taste great

(?) Now I need someone to show me what it is in Japanese :D I need it written, so symbols, not just the words If anyone has any clue, please let me know

(!) [Kat] Languages unfortunately are tied to cultures, and don't do well as simple word substitutions.
(!) [Rick] Except as a form of comedy, e.g., the gentle art of "speaking Latin like a native":
You're fired.
  Ego te demitto.
Look out, I'm going to barf!
  Cave, vomitorus sum!

Oh! Was I speaking Latin again?
  Vah! Denuone Latine loquebar?

Your fly is open.
  Braccae tuae hiant.

You're beautiful!
  Pulcher es!

Your mother is so old, she speaks Latin.
  Tua mater tam antiquior ut linguam Latine loquatur.

Show me the money.
  Monstra mihi pecuniam.

Luke, I am your father.
  Luke, sum ipse patrem te.

You don't know the power of the dark side.
  Potestatem obscuri lateris nescis.

I think Elvis is still alive.
  Credo Elvem ipsum etaim vivere.

Fuck you and the horse you rode in on.
  Futue te, et ipsum caballum.

I have a catapult. Give me all the money, or I will fling an enormous
  rock at your head.
Catapultam habeo.  Nisi pecuniam omnem mihi dabis, ad caput tuum saxum
  immane mittam.

I have a terrible hangover.
  Crapulam terriblem habeo.

Your toga is backwards.
  Tua toga suspina est.

We're off to see the wizard, the wonderful wizard of Oz.
  Imus ad magum Ozi videndum, magnum Ozi mirum mirissimum.

Only you can prevent forest fires.
  Solum potestis prohibere ignes silvarum.

I scream, you scream, we all scream for ice cream.
  Clamo, clamatis, omnes clamamus pro glace lactis.

Stand aside, plebians! I am on Imperial business!
  Recidite, plebes!  Gero rem Imperialem!

I hope we'll still be friends.
  Spero nos familiares mansuros.

God, look at the time! My wife will kill me!
  Di!  Ecce hora!  Uxor mea necabit!

You know, the Romans invented the art of love.
  Romani quidem artem amatoriam invenerunt.

Nobody dances sober unless he's insane.
  Nemo saltat sobrius nisi forte insanit.

Nu, everyone's a comedian.
  Quisque comoedum est.
(!) [Heather] For the record this made me giggle out loud:
> Oh! Was I speaking Latin again?
>   Vah! Denuone Latine loquebar?
I swear I could hear Rick saying this in my head...
(!) [Sluggo] Of course I couldn't resist translating these to Esperanto.
Ho! Cxu mi parolis latine refoje?
But more colloquially:
Ho! Cxu mi krokodilis latine refoje?
"to crocodile" comes from a character in a children's TV show. The crocodile kept babbling in his native language at Esperanto events.
(!) [Ben] I don't suppose you recall the Russian humor gazette, "Krokodil", or "Krokodil Gena" (a cartoon croc who later featured as the butt of many jokes; most of these were, unusually for Russian jokes, characterized by a certain innocent sort of humor.) For some reason, Russians seem to associate crocs with humor.
(!) [Sluggo] I've seen it. That doesn't mean I understand the jokes.
I remember one political cartoon. One side of a chessboard was all lined up orderly; the other side was squabbling. One from the unruly side goes up to the other and says, "U vas yeschyo korol' s korolevoy? A u nas uzhe yest' parliament." (You still have a king and queen? We have a parliament.)
(!) [Heather]
> Luke, I am your father.
>   Luke, sum ipse patrem te.
(!) [Sluggo] Luk, mi estas via patro.
Not "Luke" because "luko" means skylight, so "luke" means skylight-ly. "luk" alone would be a preposition or interjection, but since there's no such word it's safe for a foreign name. The accusative would be "Luk'on", the standard way to add a suffix to a non-Esperantized name.
(!) [Heather]
> You don't know the power of the dark side.
>   Potestatem obscuri lateris nescis.
That was pretty good too.
(!) [Sluggo] Vi ne scias pri la povo de la malhela flanko.
(!) [Heather]
But this combo takes the cake:
> I have a catapult. Give me all the money, or I will fling an enormous
> rock at your head.
>   Catapultam habeo.  Nisi pecuniam omnem mihi dabis, ad caput tuum saxum
>   immane mittam.
(!) [Sluggo] Mi havas katapulton. Donu al mi cxiun vian monon, aux mi jxetos grandegan rokon kontraux vian kapon.
(!) [Heather]
> I have a terrible hangover.
>   Crapulam terriblem habeo.
(!) [Sluggo] Mi havas teruran postebrion.
(!) [Heather] Final translation? Go ahead, shoot me. I won't notice, because the chariot racers of Saturn already ran me over.
(!) [Sluggo] Nu, pafu min! Ne gravos al mi, cxar la cxaraj konkursantoj jam superkuris min.
Oops, "cxaro" (chariot) looks like a nounization of "cxar" (for/because). Perhaps it slipped into the language without people noticing. Also, I substituted "it doesn't matter to me" (ne gravos al mi) for "I won't notice" (mi ne konscios) because it sounds more typical.
(!) [Heather] But how do you say: Argh, I cannot get my wireless to work! Hand me that ubuntu disc.
(!) [Sluggo] Fusx, mi ne povas funkciigi la radiofonon! Donu al mi la diskon de Ubuntu.
(!) [Ben] Yeah, I often say that - at least the first word - when I'm feeling that kind of frustration. Only you've misspelled it.
(!) [Sluggo] That's a different word: fik.
(!) [Sluggo] fusx: exclamation saying you're too scatterbrained to do something right, or something just isn't working out.
(!) [Ben] Yah, I use it that way too. A lot.
(!) [Sluggo] funkciigi: to cause to work.
The most amusing phrase is:
Ho, kio jenas en mia posxo? Cxu molaso? Diable. Dacxjeto! [1] Well, what's this in my pocket? Molasses? Damn. Davey!
jen: behold, voila. Turned into a verb.
diable: devil-ly
Dacxjeto: from "Davido" (David). "cxj" makes an affectionate nickname. "et" is a diminutive, in this case implying a child.
[1] From Being Colloquial in Esperanto by David Jordan.
(!) [Rick] As we say in California: "Habetis bona deum."
(!) [Sluggo] Perhaps look for an expression that's similar in English and Japanese rather than the closest translation to "I love you". I don't know Japanese but maybe something like "Lisa is my favorite" or "Lisa is the one". Even if it doesn't say all of "I love you", she'll remember the unspoken part every time she sees it. And when a Japanese guest sees the scroll, maybe they'll think, "I wouldn't say that but it sounds pretty cool."
(!) [Kat] While you're not far wrong in principle, I think what you're missing is that any sort of declaration like that is by its very nature (of being a public declaration) somewhat juvenile from a Japanese perspective. It's not just about the wording.
Now, doing the equivalent of building the Taj Mahal, that would translate just fine. ;)
(!) [Jimmy] "We have to include a lot of adjectives and adverbs to get across details of people's feelings, while Japanese verbs often contain these subtleties already. One result of this is that Westerners make strange blunders in Japanese because they are unaware of which emotion the verb they have chosen conveys."
Something like that?
(From "Japanese for the Western Brain", http://www.mindspring.com/~kimall/Japanese)
(!) [Kat] Sorta. Ish.
It's not that "love" has unusual subtext, it's the action of declaration is sort of...Hmm. It's like "I'm going to kiss you now". That would be weird, yes?
(!) [Jay] Not if you're going to Antioch.
(!) [Sluggo] Maybe just "Lisa" then, or "Lisa the painter" or "Lisa -- mathemetician" or whatever she's good at. Then it can look like a simple gift rather than a "declaration".
(!) [Kat] Oh, now that would be more than a simple gift, it would be very very appropriate. Even better would be getting her a seal that she could sign her work with.

Published in Issue 122 of Linux Gazette, January 2006