It’s all clear to me now

The European Commission has finally clarified the European software patent issue once and for all. Software clearly is either patentable or not patentable, depending on your opinion and on the size of your business. You can get a software patent filed at the European Patent Office, and somebody with enough money for the lawsuit can then ask the European Court of Justice to invalidate the patent on the grounds that it describes a piece of software.

So this means that dominant tech companies can add software patents to their patent portfolios in order to ensure their dominance just like in the US, and small tech companies are free to go bankrupt while fighting in court for their right to do business and innovate in Europe.

Not confused enough yet? Let Yahoo! News try and explain:
EC: Software is not patentable – Yahoo! UK & Ireland News

Technorati Tags: software patents

Open Document Format is now ISO standard 26300

The International Standards Organization has accepted the Open Document Format developed by OpenOffice.org as an international standard. Good news, freedom lovers! Perhaps governments and other stubborn bodies will not be so eager to force citizens (the owners of their own personal data) to use “industry standard software” such as Microsoft Office. Instead, as the OO.o press release states, “For the first time in the history of computing, software users will be guaranteed that they will be able to use their data in any compliant software package, both now and in the future.” Nice job, OpenOffice hackers and suits!

Linus still delivers mature, balanced arguments

In the middle of a technical debate about virtual memory in different kernels, Linus Torvalds showed again his charm and tact by claiming that “Mach people (and apparently FreeBSD) are incompetent idiots.” After Slashdot reported this, Linus wrote a nice, friendly response on his own medium (the lkml), not on Slashdot of course, pointing out that the Slashdot people “usually are smelly and eat their boogers, and have an IQ slightly lower than my daughters pet hamster”. He is in a position to say this because not only is he “the smartest person around” but also “incredibly good-looking”.

This flamebait is a worthy continuation of a long tradition of Linus’ communication, starting back in 1992 with his own master when it comes to operating system programming, and a more recent evaluation of the competence of the makers of the world’s premiere business Unix desktop.

All this is well in line with the fact that he is so smart he knows free software licensing matters better than the FSF which enabled him to write Linux in the first place, and its lawyers when it comes to the upcoming version of the GPL, the license which he licensed Linux under.

Linus fixes Kaspersky’s flawed “cross platform virus”

Kaspersky Lab's announcement of a Windows/Linux cross-platform virus was the scoop of last week (source code available of course).

Now, there are obvious difficulties to spreading viruses on Linux and other properly designed systems, as demonstrated the lack of Linux viruses despite the availability of the ELF Virus Writing HOWTO since 2002. But the saddest part is that the virus didn't actually work on Linux kernels later than 2.6.16, as demonstrated by the testing and analysis published by Hans-Werner Hilse.

Linus Torvalds agreed with Hilse's analysis but was left wondering why the virus worked on older kernels but not the post-2.6.16 ones. He examined the situation and found a bug in GCC (the GNU C compiler) which was triggered by some code in the new kernels (I'm not going to pretend I understand any of this). Naturally he was intrigued by a program which could run natively on both Windows and Linux platforms.

Linus's explanation about the bug was published today on Newsforge. The funniest thing is, Linux has fixed the flaw and made the virus work on all versions of Linux. This might come as an embarrassment to Kaspersky, who obviously was going to cash in on the Fear, Uncertainty and Doubt triggered by a Linux virus and the users' need to suddenly buy Linux anti-virus software from them. Fortunately the free software hackers were again more than happy to help make broken code work again.

Let's see if anybody will buy anti-virus software for Linux now, or Kaspersky's software for any platform.

Proprietary video drivers in Linux

Some time ago, I wrote a lament about the sorry state of video support in Linux, or more correctly, the crappy Linux support on the part of graphics chip manufacturers. Today, an interesting CNET article touches the issue. Do proprietary drivers belong in a free kernel? I’ll just summarize the views of the main manufacturers here.

ATI defends proprietary drivers on IP reasons: their drivers include licensed proprietary third-party IP that they are required to protect. They also cite teir own IP interests and wish to “maintain the proprietary, trade-secret nature of that as long as possible.” Way to go.

Nvidia doesn’t think hackers can write complicated software: “It’s so hard to write a graphics driver that open-sourcing it would not help,” Also, their customers don’t seem to be complaining enough. They do have open source “where it makes sense”. Nice.

Intel seems to be the most interesting company here. They hope to compete against ATI and Nvidia specifically with open source drivers. If their chips become good enough in the near future, I have something to recommend to newbies again!

The article also has lots of interesting comments from the Linux kernel hackers and development of a stable interface for proprietary drivers, as well as the strategies of the commercial Linux biggies Red Hat and Novell. So go ahead and read it in full.

Oracle Novell SUSE Desktop Linux Pro?

According to a Reuters scoop, Oracle wants to deliver a “full stack”, i.e. an operating system and the applications in a buldle, just like Microsoft. Linux is a natural choice for an Oracle database grid/cluster/whatever platform. That would make a lot of sense, why not? The story mentions that Oracle is thinking about buying Novell, which would mean that new Oracle systems would be built on the SUSE Enterprise Server.

Oh well, the news, especially business news is full of stories like this. Perhaps I should just go and read some real news now.

Update: Oracle has confirmed the OS plans.

SMB printing in Dapper: a bug and a workaround

People at work tossed a little Fujitsu laptop to me for some reason. It had Windows XP but I promptly replaced it with a real operating system and now it runs Ubuntu Dapper. I named it oskar, after the little hero in Grass’s The Tin Drum. Its internal network card is broken but I can replace it with a little USB LAN dongle (tested under Windows and it worked). The installation was totally uneventful, everything Just Works as advertised.

Samba printing, however, is currently broken. Because there is a Windows box on my home network, I use Samba for all networking. Samba on my Dapper server works well and the Windows box can print to it, but the Dapper laptop could not. On the Ubuntu Forums, other people seemed to be having problems with printing from Dapper to Windows servers, so i went on to examine relevant bug reports on Dapper. Looks like the problem has appeared quite recently (sometime around April 12th), and is on the client side. Sure enough reverting the smbspool binary to an older version indeed fixed the problem. Hopefully we’ll soon see this bug fixed and smbclient updated!

Otherwise, Dapper currently works very well, and the new GNOME is awesome.

Easter eggs

Warning: Easter Egg spoilers ahead!

Celebrating Easter, Jonathan has gone egg hunting. He first exposes a known liar, aptitude. We all know of APT’s Super Cow Powers because they’re documented (type “apt-get” with no arguments and hit ENTER), but aptitude keeps its powers hidden. The key to finding aptitude’s Super Cow Powers is endurance and liberal use of the -v switch.

Firefox’s secret message from the Book of Mozilla is well known as well, but the really great OpenOffice.org Easter Egg that Jonathan unveils was news to me. Of course, Jonathan sees such an elaborate Easter Egg as “proof that OpenOffice.org is bloated”. Not that we would need further proof of that! :) I didn’t know about the funny release names in the Ubuntu kernel documentation either.

Here are a couple of GNOME eggs I’m aware of. Open the “Run” dialog (press alt+F2). Type “gegls from outer space” as the command. You get to play a game that “will change the way you think of your desktop forever.” It features GEGLs (Genetically Engineered Goat, Large). GEGL is a mythical creature in GIMP and GNOME lore, and the unofficial secret logo of the GNOME project.

The other GNOME egg i know about features Wanda, the fortune telling fish from the Fish Applet. Type “free the fish” into the Run dialog and Wanda will swim around your desktop occasionally. (You can accomplish the same thing by hitting the ‘f’ key three times after opening the “About” box of a panel.) If you click it with your mouse, it will flee, only to return later. You cannot kill this process because it’s hidden in the gnome-panel process (or one of its children – killiing gnome-panel does help).

Of course, Easter Eggs in free software have given rise to some complaints as well. An OpenOffice.org user argues that a piece of free software should work as advertised and only in that manner. There are bug reports demanding the removal of Easter Eggs from OO.o, or at least an easy method for sysadmins to disable them. Issue 61685 has extensively discursive comments for and against eggs.

A picture is worth a thousand words?

In his ZDNet blog, Richard Stiennon shows maps of system calls that occur when a Web server serves a single HTML page containing a single image. He cites the thread chaos on the Windows server as proof of the operating system’s insecurity. It’s true that complicated systems offer more opportunities for crackers to utilize buffer overflow vulnerabilities compared to simpler ones. Although the images don’t really show just Windows and Linux systems but combinations of LInux/Apache and Windows/IIS, the sheer visual difference in the system call maps is stunning.

GNOME 2.14.1 hits Dapper

New GNOME versions are always first introduced to the development versions of Ubuntu (well, Foresight gives them a run for their money), so 2.14 has been in Dapper since almost day one. The GNOME project tends to release a point release pretty soon after a major release, to fix obvious bugs and stabilize the desktop for production use. I can now pronounce my GNOME deskop not only “awesome”, but also very very stable. This is clearly the best GNOME ever. I haven’t been this happy about GNOME since 1.4 or so (not that I would be very happy with 1.4 today :)

This is also a major step feature-wise: the searchable GNOME. I can have even less widgets, dangles, bells, whistles, and “stuff” on my desktop because I can very intuitively find anything without the interface getting in the way too much. Well, that’s not entirely true, but MORE true than ever before on any desktop I’ve tried. Under the hood, the GNOME hackers have really put a lot of work into making GNOME less resource hungry, so that maybe I don’t have to buy another set of hardware just because I have a new version of the software. It’s the other way around! Try telling that to Microsoft.

The last of the Linux-friendly graphics chip makers

For years, when newbies asked me which graphics card they should get when building a Linux box, the answer was easy. Unless you need bleeding-edge 3D support for, say, playing Doom 3, buy a Matrox card. I was always very happy with my Millenium G550 card’s 2D performance, which is what really counts on a serious workstation, and I always got enough DRI support to play simple 3D games and graphics hacks. Matrox used to support new versions of X with their binary drivers, which were incorporated into XFree and X.org sources pretty soon afterwards.

Now that Ubuntu Dapper uses the new X.org 7.0, i naturally wanted to play with cool and useful effects that the bleeding-edge XGL extension supports, so I needed DRI. Direct renedering was apparently not supported with the free mga drivers in X.org, so I readed for the Matrox website for drivers. Turns out Matrox had no driver for X.org later than version 6.8.1, and judging from the responses by Matrox representatives on their support forums, we shoudln’t even expect them any time soon. Perhaps the Matrox hackers have assumed the stagnated state of mind of the old XFree team and got scared of the brand new modular X that the 7.0 release represents, or perhaps the company policy has changed, I don’t know. No explanation there, beyond “no ETA at this point” for the drivers. Daunted, I went on to do other stuff.

This morning I returned to the support forums, only to find that the Linux forum was locked, no new posts could be submitted. Apparently Matrox had grown weary of the Rants of the Linux users (who had been spoiled with quick delivery of drivers in the past), and decided to simply shut off this channel of critique. Daunted, and this time very pissed too, I turned to Google for a last search for a solution.

Lo, Google directed me to Arch Linux Wiki, where somebody had indeed come up with a fix involving a simple change in the xorg.conf file and the free mga driver in X.org. I quickly applied the fix (edit xorg.conf, add

Option “OldDmaInit” “true”

to your “Device section), tested and saw that it was good.

So now I have DRI, but I have no graphics chip maker to recommend to newbies. I never guessed Matrox could afford to lose the strong support they’ve had from Linux/X users over the years now that they’re losing the game to NVidia and ATI already. Now all I can say is, ATI is the worst one because their proprietary drivers are constantly broken. Matrox is close to the bottom, since they have no binary drivers at all, and we don’t know how much they are interested in giving the X community’s free drivers. NVidia has very good proprietary drivers for linux in case you don’t mind using them.

One of the best things about Linux has been that you almost never needed to hunt down and install drivers for your hardware, as Windows users are accustomed to do. If things continue to decline, we will always need to find the correct drivers (non-free drivers no less!) to get decent graphics support.

Woe is me. Please tell me I’m wrong and Matrox cards will work out of the box like they used to Real Soon.

KeyTouch for weird multimedia keyboards

I was never able to get the default GNOME keyboard properties capplet to work with my Logitech Media Keyboard. Then I suddenly stumbled on an Ubuntu Forums post saying that KeyTouch is the “first and only program of its kind that works perfectly together with kernel 2.6”. And it does! My keyboard was even supported. You should be able to get any keyboard supported quite easily with the provided KeyTouch-editor if yours isn’t. Great work Marvin!

Linux distributions, bloggers, and other aggregators

Luis notes interesting similarities between blogging and other forms of "remixable commons-based culture":

"As long as licenses are respected, for-profit content aggregation (into a software distribution, a web-based blog aggregator, or whatever) is generally acceptable. If creators don't personally find it acceptable, the onus is likely on them to choose an appropriate license. The aggregators which are perceived to give the most back to the communities and individuals they draw from are likely to be most popular, at least among opinion leaders."

This sounds plausible, but notwithstanding hypotheses concerning the differences between Sun, Red Hat and, say, Ubuntu, proper concepts are lacking. This calls for study.

Ubuntu power

Anthony Towns has done an interesting numerical comparative analysis of Debian sarge vs. Ubuntu hoary. "For a distribution that’s under a year old to be maintaining about a quarter of Debian’s packages seems pretty impressive." Well, you could say that. How many hackers does Canonical have working on Ubuntu again? A dozen? Two?

“X Windows” information from Linspire

Dan Stone has found a fascinating piece of information about the X Window System: "Linspire's page about X is the most factually incorrect thing ever." I'm trying to think about a competing page but it sure is damn hard.

Davix takes over

Dave Jones has great news on this April 1st, announcing Davix, a free Linux-like kernel for those who "pine for the nice days of Linux-1.1, when men were men and wrote their own device drivers".

More phone suckage; newbie-love

I can use p3nfs for reading my phone now. Writing is still impossible. Nobody seems to use p3nfs anymore, and those who do, are getting perfect results. Not very encouraging, as no-one seems to have had any reason to fix the problem I'm having. Also, I can't get irda working on Linux (this one is certainly my own fault). So I'm still reduced to a Windows user when I need write support.

Helped newbies on #ubuntu, which is a nice hobby.