In the comments to the article I wrote about running the 64-bit version of Ubuntu Feisty Fawn on a Gateway MX7626, I added that my friend who owns the laptop had “upgraded” to Gutsy Gibbon Tribe 4 to try and fix a problem with intermittent sound under Feisty. The initial upgrade did work and her sound functioned properly. I talked to her again last night and she is giving up on running Ubuntu alpha software and is going back to Feisty.

The Free Software Foundation has defined Four Freedoms related to software. These freedoms apply to users of software, not necessarily developers. In the view of the FSF, these freedoms are ethical in nature, so much so that they argue that software which violates these freedoms is unethical.

Like many other rights, the four freedoms are specific expressions of abstract freedoms in the context of software. They represent concrete examples of underlying notions of freedom. You can see this principle if you ask “Why should I be able to run my own printing press or weblog?”

If those underlying principles exist, then it should be possible to identify them. It should also be possible to extrapolate concrete expressions of those principles in new contexts… such as hardware, not software.

This for the “how the hell have I done this job this long & not known this already?” files.

Debian has a file called /etc/rc.local which runs at the end of all the multi-user boot levels, and which you can therefore put stuff in. I’ve had trouble with autofs not starting properly on certain machines (there seems to be a correlation with SCSI or SATA rather than IDE drives, although I do not know why this should be), and putting the line

/etc/init.d/autofs restart

Especially baffled that I didn’t know of this file because I’ve had to do stuff like this in the past, and have horribly misused /etc/init.d/rmnologin instead.

In totally unconnected, and not even slightly tech-related news: a bunch of my cycling friends are riding 1200km in 90hrs in France this week, for fun. (I was hoping to go as well but didn’t qualify - although given the vile weather they’ve had this might be for the best.) They’re all doing fantastically well, especially given the awful conditions, & I’ve been following their progress all week with much excitement (one person has finished already, in 60hrs!). Finish deadline is 4pm French time tomorrow. Allez allez!

This press release might be of interest to some folks:

“LPI exams to be offered at half-price at Linuxworld Conference and Expo 2007. Industry exams are half price to all conference delegates and show attendees.”

Back in my January review of Vector Linux 5.8 Standard, the version with the Xfce desktop, I touted Vector Linux as the fastest distro with a reasonable feature set and selection of software that I had used at the time. It took a while but I finally found a distribution that’s at least Vector’s equal: Wolvix 1.1.0.

I’ve tested Wolvix on two laptops: a not quite two year old Gateway MX2676 (AMD Athlon 4000+ mobile processor, 1GB RAM) and a nearly five year old Toshiba Satellite 1805-S204 (1 GHz Celeron processor, 512MB RAM). The Gateway’s processor is 64-bit but Wolvix, at least so far, is only available for 32-bit x86 architecture. Performance was impressive on both machines but Wolvix truly shined on the older Toshiba.

Much like Ubuntu, Wolvix is provided as a single iso image of a Live CD with a graphical installer. Wolvix also offers the option to run entirely cached in RAM provided you have enough memory. Wolvix also offers a frugal install where the iso image is installed directly to the hard drive and is booted read-only. You are then effectively running the Live CD with the speed of a conventional hard drive. With four different ways it can be run Wolvix is a very flexible animal indeed, a distro which can be easily tailored to a number of specialized uses and yet is still brilliant as an ordinary distro installed to your hard drive.

Wolvix 1.1.0 is a user friendly distro based on Slackware. The code base appears to be a heavily updated Slackware 11 rather than the current Slackware 12. Previous versions were actually remasters of Slax, a small live CD built from Slackware with the Linux Live scripts, but with version 1.1.0 Wolvix has struck out on it’s own path and is no longer built from Slax. The improvements since version 1.0.5 are dramatic. While it’s still not perfect by any means the new version of Wolvix is relatively user friendly and easy to use.

This just keeps getting worse. Please tell me how anyone could think this is a good idea:

Tech Tips with Gnull and Voyd
http://www.linuxjournal.com/article/9452

“Howdy. My husband is Chester Gnull and I’m Laverta Voyd, and I’m the lady to light a way for all you sweethearts out there who do fancy stuff with Linux. Me and my husband’s gonna be bringing you tech tips just about every month now….I don’t know nothing about Linux. Chester, he’s the smart one…”

How many moronic stereotypes can you count in this? I thought the odious Mango Parfait http://www.tuxmagazine.com/node/1000131 in Tux Magazine was bad; somehow the fine folks at Linux Journal managed to exceed their previous Standard of Awfulness.

Here is a free suggestion for the fine folks at Linux Journal, though given the systemic awfulness of their perception of women it’s probably wasted: women don’t need husbands and boyfriends to do tech. We don’t need to be portrayed as cartoony stereotypes. We stand on our own merits as competent, skilled adults and tech professionals. Are you capable of understanding this? Why do you have an editorial policy of belittling and being disrespectful to women?

When are you going to issue a public, sincere apology? When are you going to get a clue and quit portraying women in such idiotic ways?

Part 1:
http://www.oreillynet.com/linux/blog/2007/08/linux_journal_the_last_idiots_1.html

Linux Journal contacts page: http://www.linuxjournal.com/xstatic/staff/index

I use RT as a request/bug tracker, but until recently hadn’t set it up with an email address plugged directly into it. This was because I don’t run my own email server - that’s centralised - which makes setup a bit more difficult. And undocumented, hence this post. Convincing users to use a different email address may well be tough, but at least you yourself can start bouncing relevant emails to the RT address, thereby creating a more trackable system.

There are 2 basic steps: 1. setting up the mail gateway to RT; 2. mail pickup from the external central server. Note that I’m using exim4 - other mail programs will obviously work differently. Details are below…

I dropped my subscription to Linux Journal a couple of years ago because I got tired of their stubborn refusal to admit women into their gearhead club. Sure, you’ll find the very occasional women author, but not many. What I did see a lot of was pages of ads that insulted women in every stereotyped way- brainless ornamental boy toys. God forbid we should ever be portrayed as IT professionals and actual humans.

Fast-forward a year or two, and what happens? Does LJ wise up? Why, no! Not at all! Instead they take a huge step backwards and run a full-page ad that insults both women and men:
http://feministing.com/archives/007501.html

Way to go, Linux Journal! When can we expect your sincere, very public apology? When are you going to wise up?

Some more links:
http://www.allbusiness.com/marketing-advertising/4187845-1.html
http://use.perl.org/~brian_d_foy/journal/33953?from=rss

Linux Journal contacts page: http://www.linuxjournal.com/xstatic/staff/index

Part 2 “Dear Linux Journal: News Flash- Women Are People”
http://www.oreillynet.com/linux/blog/2007/08/dear_linux_journal_news_flash_1.html

**edit**
I changed the title from “Idiot-Boy’s Club” because it was unfair and mean.

It is two days after VMWare had one of the most successful IPOs on recent years, and one day after XenSource announces that it is being acquired by Citrix. Now that money is flowing into the two major virtualization players at a rate we haven’t yet seen, what are your feelings about using Xen or VMWare on Linux? Which are you considering running on your next production network?

Update 8/16 2:28 PM (Central): From the comments: Xen: 1 vs. VMWare: 0 vs. Neither (libvirt): 1

Articles about Linux and mainstream Linux news tends to be dominated by the big Linux distributions, those with large corporate backing and/or large development teams. I’m primarily talking about Red Hat Enterprise Linux and it’s free clone CentOS, Novell/SuSe, and Ubuntu on enterprise servers and Ubuntu, Fedora, Linspire, and Mandriva on the desktop. Throw in two venerable and widely respected distributions, Debian and Slackware, and you’ve got about 90% of the industry chatter covered, maybe more.

These distributions also have something else in common: with the exception of Linspire/Freespire, which I haven’t tried, they have all frustrated me on one level or another. I’ve found recent Fedora, SuSe, Mandriva, and Ubuntu releases all to have more bugs than I would expect, often very annoying and obvious ones. All of the above mentioned distors except Slackware are unimpressive in terms of performance. Most tend to be bloated and full of all sorts of cruft that I don’t need that gets installed by default. The notable exceptions are Ubuntu and Mandriva One which both come on a single CD and install a stripped down, clean OS which you can then build on. However, in the case of both Ubuntu and Mandriva One they seem to get a whole lot less useful stuff on that single CD than some other distributions seem to manage.

In the past year a number of medium sized and small distros have leaped past the big players among Linux distributions, offering single CDs with lots of apps, excellent hardware support, speedy performance, and relatively few bugs. Some are also far more user friendly than distros like Ubuntu and Mandriva, often touted as the best place for a newcomer to Linux to start. When I say medium or small I’m referring to both the developer community and user community around each distro. In some cases the developer community is just one or two people.

run-parts is used (on Debian systems, anyway) to run the scripts in /etc/cron.daily (hourly, weekly, etc) on the appropriate schedule. I had trouble this week with a Perl script I’d dropped into /etc/cron.daily failing to run. Ran fine from the command line, of course. Odd.

Eventually it occurred to me, after a little light man page reading, to try run-parts --test /etc/cron.daily (which just prints the names of the scripts that would run). Script failed to show up. Most Odd.

I finally found the answer via Google, although a slightly less light reading of the man page would have helped. Scripts to be run by run-parts must adhere to a particular naming convention - in particular, no .xx endings. So my script.pl script wasn’t being picked up due to that .pl ending. I renamed it to script and all was well.

(I’m not actually sure what the logic of this is; I’m assuming it’s likely to be historical reasons. You can alter it with the --lsbsysinit option, if you prefer that. I know the .xx ending is by no means essential, but I prefer in general to have a quick visual of what language I’ve written a script in.)

Over the past couple of weeks I’ve been able to borrow a friend’s laptop: a Gateway Over the past couple of weeks I’ve been able to borrow a friend’s laptop: a Gateway MX7626, model W730-K8X (Athlon Mobile 4000+ processor, ATI X600 graphics, 1024 RAM). She has the 64-bit version of Ubuntu Feisty Fawn installed. I’ve installed and worked with 64-bit Linux on servers over the past couple of years, mainly running Red Hat Enterprise Linux 4, but this is my first chance to play with 64-bit Linux extensively on a laptop.

First, the machine is wonderfully fast at everything I’ve tried to do with it. 64-bit Ubuntu does have some minor quirks. The most noticeable one is that sometimes sound works and sometimes it doesn’t. If I don’t hear anything when GNOME starts then I won’t have sound until I reboot. She obviously has ALSA configured correctly since there is sound more often than not. I also noticed that some graphical apps don’t have .desktop files in /usr/share/applications and consequently don’t show up in the menu. When it comes to anything truly important, though, 64-bit Feisty does seem to work very well.

I attended my first OSCON two weeks ago. I learned a lot about leading open source projects — something that has long been important to me, but now is even more important for me, since I’ve just started working full time on Intel’s new Threading Building Blocks open source project. I split my time at OSCON between sessions related to multithreaded programming and open source project management strategies.

It’s easy to think that an open source project’s success depends on the quality of the product and nothing else. But that is surely not the case. An excellent open source project that no one knows about is very likely to eventually disappear.

Yet, open source projects are indeed quite different from corporate products. An open source project succeeds through the interest and effort of its community. It’s not just a matter of how many dollars were spent to promote the product — what matters as much, or more, is the community, the relationships, the dialog among members of the community.

A given is that the open source project must be worthy of attention, the software must address a need and be of high quality. Otherwise, why should anyone pay attention to it?

Why corporations support open source projects

In his recent O’Reilly Radar post “Yahoo!’s bet on Hadoop”, Tim O’Reilly commented:

OK — but why is Yahoo!’s involvement so important? First, it indicates a kind of competitive tipping point in Web 2.0, where a large company that is a strong #2 in a space (search) realizes that open source is a great competitive weapon against their dominant competitor. It’s very much the same reason why IBM got behind Eclipse, as a way of getting competitive advantage against Sun in the Java market. (If you thought they were doing it out of the goodness of their hearts rather than clear-sighted business logic, think again.) If Yahoo! is realizing that open source is an important part of their competitive strategy, you can be sure that other big Web 2.0 companies will follow. In particular, expect support of open source projects that implement software that Google treats as proprietary.

So, corporations support open source in order to make money. Duh! But, let’s think about who benefits: shareholders, employees, the company’s customers, and whoever uses the supported open source software. So, in the case of Yahoo! or IBM, or Sun, or Intel — is there anyone reading this post who isn’t among that group of beneficiaries?

Corporate reticence

To be a good open source citizen, it seems to me that a corporation needs to understand the value of reticence. Corporate reticence implies that the corporation understands where the world is moving, and they want to ensure that they have their boat in that rising water. It’s a statement that they don’t employ the “not-invented-here” syndrome in their decision-making. A statement that they would like to be a member of the community that is creating the future, rather than the dictatorial creator of that future.

This is an enormous step for corporations to take. Open source has played a great role in bringing this situation into being, in my view. The notion that a free community can be more powerful than a corporate spreadsheet is difficult to swallow for most corporate officers. Yet, some come to understand it much more quickly than others.

OSCON and Corporations

Corporations, of course, are the prime funders of conferences such as OSCON. It’s a signal that corporations (some of them, anyway) have recognized that the world is much bigger than their specific product domain. And if they can be good citizens within that larger world, they will reap benefits along with all other participants.

There is definitely a big difference in how well corporations understand the open source process, and how they address “openness” in general. In my view, some — such as IBM, Intel, and probably Sun, “get it.” They simply want to position their boat in the rising water, and to do so they are willing to pull their fair share of the weight. If open source projects have hundreds of salaried IBM, Intel, and Sun developers working on them, isn’t that a great benefit? I’d say so!

I was able to attend OSCON because I’m working with Intel on Threading Building Blocks. It’s a project I’m very pleased to be working on, because much of my software development experience has involved multithreaded software development, on both Unix and Windows systems. I thoroughly enjoyed OSCON. I’m excited to be working on my first open source project, Threading Building Blocks. While Intel invented the software, the intent is indeed that its future will be determined by the community. That’s the kind of corporate participation in open source that I see as beneficial for the entire community.

At OSCON I saw evidence that many other corporations are starting to “get it.” I consider that a very very good sign!

I just ran across a great tutorial by Miriam Ruiz: Recipe: Internationalizing a C program with gettext. Though it’s only a starting point, it explains everything I failed to get by trawling the appropriate manpages. Now I have almost no excuse not to do it!

It appears that Vector Linux is moving to a faster release schedule. Vector Linux SOHO 5.8 is only a few months old but the first release candidate of Vector Linux SOHO 5.8.6 has been announced:

VectorLinux is pleased to announce the availability of the rc-1 release of 5.8.6 for the VectorLinux SOHO product. This RC release includes the following improvements: upgrade to Xorg-7.2 from the previous 6.9 series which adds many gui improvements and 3D enhancements: rebuilt many core packages so they install properly in new Xorg-7.2 file structure: rebuilt the core font subsystem to take advantage of the new Xorg release: Updated many packages such as pidgin (including spellcheck) and Xscreensaver…, general bug fixing and driver updates. Beryl is an installable option for those wishing the ultimate 3D desktop experience. There are many other enhancements to numerous to mention.

I think this is the first time they’ve had a release of the KDE-based SOHO without a Standard or Deluxe release. This appears to be part of a general change in direction for Vector that’s been happening for the last year. They seem to have moved well away from being a specialty distro for older hardware to being a full featured, general purpose distro. The recent move to begin selling commercial support echoes that transition as does their recent appeal for volunteer testers and developers for a future 64-bit version. I see this shift as entirely positive provided the lightweight Standard version continues to be developed and supported.

In recent comments to my review of Vector Linux 5.8 SOHO keyfitter wrote:

There is a reason why they are using Win 98 in 2007. I think it’s called, being cheap!.

I wonder if these people realize they can buy a brand new computer for $139. Granted the hardware is a bit dated by today’s standards but it’s probably light years ahead of what they are running Windows ‘98 on. Of course these computers come preloaded with Linux: Vector Linux 5.8 Standard to be precise. That’s fine. Without having to worry about installation or hardware compatibility someone who buys this system gets a nice, ready to go, user friendly Linux system with a warranty. The return policy is listed as “no matter what” short of physical abuse. What they don’t get are Windows virii and malware. They do, of course, have to learn a new OS.

Why not offer the same system with Windows? It would nearly double the price. People forget that they pay an average of around $100 for the privilege of having Windows on their new computer. Of course Windows Vista wouldn’t run on a 1.5GHZ system with only 256MB of RAM, would it? Vector Linux 5.8 Standard will run quite nicely, though. A memory upgrade wouldn’t hurt particularly if you’re interested in a lot of multimedia applications (an extra $39 for 512MB) but it isn’t strictly necessary.

I’ve had Google Desktop for Linux running for a while now; I can’t actually say I’ve used it all that much, but what I did very quickly observe was that the default cronjob (/etc/cron.hourly/gdl-update) that checks for updates had far too much output (I do not want hourly updates telling me “not going to update now”, kthx). And some spelling mistakes.

Here’s an improved version that only sends you email if something of note has happened (update occurred; update failed unexpectedly; can’t create or update timestamp file):

#!/bin/bash
# Copyright 2007 Google Inc. All Rights Reserved.

CACHE_DIR="/var/cache/google/desktop"
PREFIX="/opt/google/desktop"
GDL_UPDATE="/opt/google/desktop/bin/gdl_update"
PKG_FORMAT="deb"
PKG_UPGRADE_CMD="dpkg -i --refuse-downgrade"
TIMESTAMP_FILE="/var/cache/google/desktop/update_timestamp"
ID_FILE="/var/cache/google/desktop/id"
PATH=/sbin:/usr/sbin:/bin:/usr/bin:$PATH

if [ ! -x "$GDL_UPDATE" ]; then
  echo "gdl_update is not available."
  exit 1
fi

# run gdl_update
export PATH

DO_UPDATE=no
if [ ! -e $TIMESTAMP_FILE ]; then
  if touch $TIMESTAMP_FILE; then
    DO_UPDATE=yes
  else
    echo "Failed to create timestamp file."
  fi
else
  LAST=`date -u -r $TIMESTAMP_FILE +%s`
  NOW=`date -u +%s`
  DELAY=`expr 86400 + $RANDOM % 21600`
  DIFF=`expr $NOW - $LAST`
  # Update should only be occurred every 24 to 30 (randomly) hours.
  if [ $DIFF -gt $DELAY ]; then
    if touch $TIMESTAMP_FILE; then
      DO_UPDATE=yes
    else
      echo "Failed to update timestamp file."
    fi
  fi
fi

if [ "$DO_UPDATE" != "yes" ]; then
  exit 0
fi

UUID=""
if [ ! -f "$ID_FILE" ]; then
  UUIDGEN=`which uuidgen`
  if [ -n "$UUIDGEN" -a -x "$UUIDGEN" ]; then
    $UUIDGEN > $ID_FILE
    chmod 644 $ID_FILE
  fi
fi
UUID=`cat $ID_FILE`

if [ "$UUID" = "" ]; then
  UUID="0"
fi

NEW_PKG=`LD_LIBRARY_PATH="/opt/google/desktop/lib:$LD_LIBRARY_PATH" $GDL_UPDATE "$PKG_FORMAT" "$UUID" 2>&1`

# Note spelling error (appears to be in gdl-update binary so can't readily be removed)
if [ "$NEW_PKG" = "unavaliable" ]; then
    # echo "Package is not available
    exit 0
fi

# update package is successfully downloaded
if [ $? -eq 0 ] && [ -f "$NEW_PKG" ]; then
  dpkg -i --refuse-downgrade "$NEW_PKG"
  if [ $? -eq 0 ]; then
    echo "Updated successfully."
  else
    echo "Update failed."
  fi
  rm -f "$NEW_PKG"
fi