Older blog entries for mbanck (starting at number 31)

High time for some Hurd updates I say.

Since FOSDEM, a lot of things happened, except Marcus finishing SysV shared memory support... However, he intensively studied and discussed kernel capabilities with the L4 people in order to improve the Hurd/L4 framework and recently returned and submitted a patch for proxy memory objects in GNU Mach, which is one of the prerequisites for shared memory, so hopefully this will be sorted out soon.

Neal was incredibly productive, he first finished the second part of his Hurd/L4 virtual memory framework and then took a while off Hurd/L4 in favour of improving Hurd/Mach. He fixed a couple of test-suite failures in libpthread, implemented POSIX semaphores, came up with a Mach alloc tuning patch which improved overall system stability quite a lot and then fixed pflocal so that the orbit test suite passes and gamin does not freeze the whole system on startup. He further hacked a lot on gamin, fixing a couple of bugs and most importantly contributing a native Hurd back-end. He also helped with porting all over the place.

Other upstream hacking included some great work done by Gianluca on unionfs, implementing write support. He also managed to get oskit drivers running on GNU Mach-1.x, which might improve hardware support considerably. Marco did not have much time for Hurd hacking over the last months, but now he appears to be back after he graduated from school and committed to writing a native TCP/IP stack over summer. For Hurd/L4, Mathieu Lemerre ported a simple IDE driver, so that port will probably pick up speed again soon. He is also working on memory allocation.

As for porting, I managed to get GNOME mostly working, although it is quite a bit more sluggish than on Linux and not all applications are ported yet. But Neal's work on the Hurd's pflocal, orbit and gamin made it at least possible to fully start up GNOME (screenshot). The remaining patching I had to do were due to minor portability issues or disabling functionality the Hurd does not yet have (mostly sound/gstreamer). Since the upload of GNOME-2.10 into unstable, we started uploading GNOME packages to ftp.debian.org directly as bugs are now getting fixed rather fast.

I also managed to build Qt packages and run Qt-Designer successfully. I announced their availability, but nothing happened for some time until Oleksandr Shneyder stepped up declaring his intention to port KDE to the Hurd. He had some initial problems but then managed to port kdelibs and then the whole of KDE in a couple of days (screenshot).

Last but not least, Colin Watson started looking into porting Debian-Installer to GNU/Hurd, an area we quite lacked at, as the K-series ISOs are still using potato's boot-floppies and crosshurd is a mere hack (though a very nice one). Hopefully this will mean that it will get increasingly easy to install Debian GNU/Hurd soon. Due to his efforts, we now also got a working OpenSSH again, which even has X11-forwarding fixed.

Munich chooses Debian, talking to the people involved

Although most people suspected SuSE would win the europe-wide bidding for the City of Munich's migration from Windows to Linux, it got announced last week that two local german companies using Debian as the base for their client have been awarded the tender. The bigger company of the two, SOFTCON, is located in Munich and has around 120 employees. It is 20 years old, and thus not a Linux-specific company, but rather a general IT service provider. The other company (GONICUS) is located in Arnsberg and much smaller (15 employees) and more technically oriented on the migration to Linux (it also a member of the Linux-Verband). A quick googling for its name on debian.org reveals that GONICUS employs at least one Debian Developer, Cajus Pollmeyer (cajus at debian.org), so they are giving back a lot. Congratulations to everybody involved.

As it turned out, we rather accidently had one of the irregular Debian meetings in Munich organized for yesterday, so I decided to invite the people from the city hall involved with the so-called LiMux project to join us in light of them choosing Debian, and they did.

This time, we did not meet at our usual location (the Augustinerkeller), but rather at the Pasta & Basta at the Münchner Freiheit. We ordered a table for ten people, but I counted 18 people attending, so we had some logistic problems (and the waiter really was "italian-simpatico-rude" as one of the comments at the above link indicates). Out of those 18, three of them were from the LiMux project and I had a nice chat with them. They were in their beginning-to-mid twenties and really nice and friendlypeople. These are some of the points they told us:

  • They are a group of around a dozen people, some of them working for the city for several years already, some of them being employed since the beginning of the year. They seem to be determined to largely conduct the migration themselves, and only rely on the two above mentioned companies for initial bootstrapping and training.

  • Contrary to what I thought, the servers were never at stake, the migration only covers clients, as they do not have any Microsoft servers in the first place.

  • They do not have a complete picture of how the client is going to look yet (e.g., will it include OpenOffice 1.x or 2.0?), but they hope to migrate the first productions systems by the end of the year. Partial migrations like switching people from Microsoft Office to OpenOffice while still letting them keep Windows temporarily are also considered.

  • The biggest hurdle (besides training the personnel) is the large variety of custom Windows software for individual departments. They said they wrote to almost 300 software companies they got programs from about their plans with Linux, but only a tiny fraction of them responded, and even less saying Linux was on their radar. Things seems to be getting better though, as a couple of companies have contacted them lately. In general, they try to get as much as possible done via platform-independent web clients.

  • Despite the sometimes slightly bad public image of bavarian politics, the process seemed to have been conducted very objectively. They said each submission got scored by a lot of factors, and finally the one with the best overall quotient got chosen. However, they also mentioned that they had to slightly steer the negotiations subtily from time to time, as the less tech-savy city officials were sometimes deluded by marketing speak. All in all, it really sounded that the best contender really won.

  • The software patents issue dragged them down for longer than I thought, as it took them a while to get legal advice.

Besides the LiMux people and the usual suspects, there were also a couple of other people attending, most notably a guy who built the world's smallest Linux box (he told me he even got a mail from NASA saying something literally similar to 'This is Houston, Mission Control. We like your device.' after being slashdotted), Murray Cumming (of GNOME fame) and Holger Blasum from the FFII (who reported that the crucial second reading of the software patents directive in the European Parliament will most likely take place in early July and that former french minister Michel Rocard has published a good position statement on the software patents directive).

In summary, it was a really nice meeting (or rather, Stammtisch) again, we just have to manage the growing popularity (probably by moving outside to a Biergarten next time).

9 Apr 2005 (updated 9 Apr 2005 at 18:39 UTC) »

So it is the time of the year when the next Debian Project Leader is being elected. This year, it took me quite a while to vote, partly due to being busy at the University during the day and doing other stuff in the evenings, partly because there were more alternatives this year and the choice was tough.

Figuring out that Jonathan Walther would be last and `None of the Above' second-last was easy enough, but ranking the others proved to be a bit more difficult.

In fact, I still think that our current DPL, Martin Michlmayr, is pretty much the perfect person for that position, despite him being somewhat invisible as DPL during 2004: He knows a lot of Developers personally, and has good personal relations with most of the key developers; he knows how Debian works internally better than about anybody else, including spotting and trying to fix problems or motivate the appropriate people to do so (although this is sometimes are hard job) and put a lot of work into the job which was probably not visable from the outside. When I heard that he was not up for reelection, I hoped Jeff Bailey would run for DPL, but he understandably had to drop any possible ambitions once he got employed by Canonical.

I don't know Angus Lees personally (or otherwise much, really) and Branden Robinson just acts too much like a politician these days, though he was the first individual I recognized in Debian due to his witty (and sometimes sarcastic or inflammatory) remarks on debian-devel back then. Between Matthew Garrett, Andreas Schuldei and Anthony Towns, the choice was pretty tough. They all complement each other in some way and all have their different strong points and perhaps some minor weak points, so something like a blend out of them would be perfect, or all three together as a team. Speaking of which, I think Project Scud is a great idea (other projects don't have a single leader, either) and the members are well chosen.

In the end, I guess the fact that I believe Matthew has just the right balance of being cool, humorous and reasonable tipped the scales and made me vote for him first and Andreas and Anthony tied at second.


Wow, this one rocked even more than last two. Martin and Dogi weren't around this time, so I was a bit worried at first. But in the end, hanging out with the Hurd hackers like Marcus or Neal was as much fun as expected.


The saga started on Thursday evening at a subway station in Garching near Munich, where I picked up Holger Blasum from the FFII and a random girl which saw my note at mitfahrgelegenheit.de. It turned out that the FFII had just gotten a warning about the EU ministry council wanting to put their version of the software patent directive on Monday's fishery council agenda. During the trip to Frankfurt, Holger vigorously phoned people from all over europe trying to get a confirmation on the issue and also to rally support against this looming agenda item.(Damn. One week later, the ministry council has acked its software patents directive) We arrived at Frankfurt around midnight and while Holger slept at a fellow FFII supporter's place, I spent the night at my parent's house.


I picked up Holger again in Frankfurt in the morning (a little later than I planned) and after some minor navigation errors we were on the way to Brussels. We made a stop in Jülich to visit the Credativ crowd in their office. We arrived just when they had their lunch break, amu (Andreas Müller) had given me directions the previous night and it was pretty easy to find. They are located in the 'Technologiezentrum' vaguely outside of Jülich and have half a dozen of nice bureaus. I talked a bit with Noel and amu and then later had an extensive conversation with Michael Meskes about the current state of Debian with regard to Ubuntu and some commercial aspects of the Free Software world, which was interesting as usual.

We left Jülich at early afternoon and arrived in Brussels at around 5 PM. I dropped off Holger at the university and then tried to find the appartment. It took me some time, but after asking a couple of people on the streets and some extensive studying of a map in a grocery store, I finally found the guys. Neal and Barry had organized two connected appartments only a couple of kms away from the university where FOSDEM takes place. They were really nice and we had a small kitchen (no oven though), a fridge and a dish washer. There were almost enough beds available for everybody, I managed to be able to sleep in one of those.

When I arrived, Sören Schulze (sdschulze), Neal, Marcus, Marco Gerards, Bas Wijnen (shevek), Barry deFreese, Olaf Buddenhagen, Guillem Jover (braindmg) and Ognyan Kulev (ogi) were already there (from left to right, ogi is not on this picture), Jeroen Dekkers (who stayed elsewhere) and Robert Millan (nyu) joined us later on, as well as Neal's wife Isabel. By the time I had unpacked my things, fired up my notebook and read the remaining mails from that morning, the guys had managed to get an outward ssh connection on port 53 through the hotel's WLAN, i.e. free internet for the rest of the weekend...

Neal (with some small help by Barry and me) prepared a very nice dinner and then we hacked for the rest of the night. Barry managed to get Marcus to work on shared memory instead of his slides, and I mostly entertained myself with exploring the possibilities of starting the Hurd console on bootup and looking into using sbuild in a chroot, which turned out to be perfectly usable once you work around a bug in sudo. Totally exhausted, I went to bed at some point between 4 and 5 AM.


Getting up was pretty hard, but breakfast was nice. While the others left for Richard Stallman's keynote, Neal, Marcus and I went shopping in a nearby supermarket to provide pick-nick style lunch for the Hurd developer's room. At about the same time we finally arrived at devroom, the first people started popping in. It turned out the FOSDEM program had a different opinion on our schedule than Neal had, so we had to tell a couple of people to come back later.

Ogi kicked off the devroom with his talk (slides) about extending ext2fs beyond the legendary 2 GB limit. He explained the limitations of the old approach and in which way he modified libpager. He also mentioned his work on ext3fs.

Neal then gave a charismatic and enthusiastic presentation (slides) about the problems of Hurd/Mach and how Hurd/L4 is going to address them. He was very good at conveying how applications should be given the possibility to page themselves and how it is impossible for the kernel to guess the right eviction scheme. He was repeatingly emphasizing his words with his gestures when he talked about pushing stuff out of the kernel and he kept on smiling when he explained his reasons and plans.

    <racin> neal: your code seemed great at a first glance, but now that I see your slides, it seems even greater :)

Marcus was next, he talked (slides) about inter-process communication (IPC) in the Hurd/L4 multi-server context, what different kinds of IPC there will be and what security implications have been considered for the interaction between untrusted servers. He also compared Mach's RPC (very (too) featureful, but slow) to L4's (very basic, but extremely fast) and had some nice pictures to get his points across.

Peter de Shrijver (p2) continued by talking (slides) about the proposed device driver framework for Hurd/L4. He explained the differences between different bridges like PCI or USB, how the bridge drivers would interact with the device drivers and what the interrupt handler would look like.

Marco was last and presented (slides) GRUB2, the next generation all-purpose operating system^W^Wboot loader. They seem to have come a long way and it looks like GRUB2 will be easy to hack on, he cited a couple of code snippets and interfaces to prove this.

All in all, the devroom was really crowded most of the time and the talks seemed to be well received by the audience. After the talks, we finally got a chance to chat with everybody, especially the french (Debian) GNU/Hurd hackers from the HurdFr organization like Manuel Menal, Marc Dequènes (Duck), Gaël Le Mignot (kilobug) and Arnaud Fontaine. Some other people like Yoshinori Okuji (The GNU GRUB maintainer) and Christopher Bodenstein (Physicman, who is helping with the Debian GNU/Hurd port) were around as well and we had an ad-hoc keysigning party. I also met Dafydd Harries for the first time in real life, who shortly visited the Hurd developer room.

We went home to the appartment, which was a bit more difficult than expected, because we could not find the official exit of the FOSDEM parking lot and (like always) just used the entrance. On the way back, Bas voiced his interest in Debian development and Guillem and me explained the necessary steps to participate and what one can do until one has an account. Hopefully, this will mean one more very clueful Debian new maintainer candidate. We only stayed at the appartment for a short while and then met again with the HurdFr crowd at a restaurant nearby to have dinner. I talked a lot to Duck (Marc Dequènes) about Debian, the Hurd and Ubuntu and also to Ogi about his plans in Debian development since he has entered the new maintainer queue. Seems like he wants to maintain his ext3fs work as a Debian package and might also work on porting debian-installer (he is already translating it into Bulgarian), which would be great.

After dinner, we went back to the appartment for some night hacking. Marcus and Barry cooked up a glibc patch necessary for shared memory support and discussed it with Roland via mail. I also had some nice discussions with Neal, Marcus and Guillem about the general direction of Debian's GNU/Hurd port, and contrary to my previous beliefs they still seem to be interested in the Debian port and also acknowledge that Hurd/Mach is still very important today until Hurd/L4 is ready. I spent the rest of the night finalizing the slides for my 'Debian GNU/Hurd' talk I was supposed to deliver the next morning.


Getting up as early as 8:40 AM was even worse than the day before, but with some luck and tough driving we managed to be only 5 minutes late for my talk. However, when we arrived in the Debian developers room, there was no beamer available yet, and I had to start my talk without one. Wouter told me the beamer should arrive any minute, so I decided to leave my notebook turned off to still catch the 'Ohh, XFCE4 runs on the Hurd!' reactions when it would boot up. It turned out that the beamer took longer to arrive than expected and I quickly ran out of things to say from the top of my head, so I had to switch on my notebook nevertheless and look at my slides. This all resulted in the talk being pretty unorganized, but there were a couple of questions afterwards and I was moderately happy how it turned out in the end. (A couple of days later, Wouter popped up on IRC and told me he reinstalled the Hurd, so there was some immediate success from my talk) Right after me, Guillem talked about the porting issues we face and how to prevent them, by not targetting GNU/Linux but POSIX. As a lot of Debian developers were around, I hope his talk had helped to open their mind to think beyond GNU/Linux.

I stayed in order to listen to Hanna Wallach talk about debian-women and Matthew Garrett discuss the Debian Free Software Guidelines. The devroom was packed during Hanna's talk and she did a great job in communicating the aims of the debian-women project to the audience. She also mentioned she recently joined the Debian New Maintainer process, yay. I first met her two years ago at FOSDEM together with Matthew and back then was under the impression she was some sort of BSD hacker too cool for Debian or something, so I was a bit surprised when she mentioned simply having been too intimidated to start joining Debian. I guess the public image of the NM process really needs fixing... Matthew's talk was very interesting but also a bit sad, as the bottom line was that different parts of the developer body have diametrically opposed opinions on almost all aspects of the DFSG and consensus on the current DFSG is impossible (and fixing probably very hard).

After the talks I went back to the info stand with Robert Lemmen (who came in during my talk) and met the others to have lunch. Unfortunately, we did not find a suitable restaurant and had to cope with the sandwiches sold at FOSDEM. Afterwards, we met a couple of Debian-UK guys in the big auditorium and hat some more ad-hoc keysigning there and then listened to Alan Cox talk about stable kernel development. The bit about 'Once you fix the VM for one use case, it breaks for another' part was funny in the light of Neal's talk the day before, where he identified exactly this as a fundamental problem and pointed out how to address it by moving the memory management under the control of the applications themselves. I later listened to the second half of Thomas Langes' FAI talk and then went back to the appartment with Marcus, Olaf, Guillem and Ognyan to pack. When we came back, I met Robert again at the Debian booth and we listened to the second half of the GPL enforcement talk. After that was over, we realized a lot of people had left already. We could not find anybody to go out for dinner with and thus decided to drive home early.

We picked up Paul Sladen and a desktop box, both of which the german guys "forgot" at FOSDEM. I mostly talked to Robert on the way home, he seemed to be interested in the Hurd, so maybe a new developer is born. We also talked about other aspects of Free Software, from Debian over Ubuntu to Java stuff. When we arrived in Frankfurt, it was already pretty late and Paul couldn't reach his contact in Marburg, so he had to sleep at the same FFII activist Holger slept three days earlier. I decided to drive all the way to Munich, which was pretty exhausting and Robert had to tell me lots of different stories to keep me awake. At some point around Würzburg I suddenly realized we still had Martin's desktop box in the trunk, but it was already too late to turn back.

Finally, at 4 AM, I was back home from a blast weekend (and recovering ever since. I met Martin two days later near Nuremberg and handed over his box.)

23 Feb 2005 (updated 24 Feb 2005 at 15:08 UTC) »

My Thinkpad's harddisk started to spew out errors in early December, and it took me some time to, ehm, update my backups and get a new harddisk shipped. When the new disk finally arrived, I decided to install Ubuntu on it; partly due to a Warty CD being a bit easier to reach than a Debian one, partly because I don't like having to remove all the KDE applications after doing a Sarge desktop install.

Over the last couple of months, I became co-maintainer of some packages, namely exult (an Ultima7 game engine, I have since taken over maintainership completely), sbuild (the version in the archive directed for end users), crosshurd (a hackish way to bootstrap a Debian system (Linux, BSD or Hurd) using parts of debootstrap and apt-get) and perhaps most notably the hurd package itself. Since I started my Ph.D. in November, I had to cut down on time devoted to Debian and mostly looked after the hurd package. The most important change here was the inclusion of Ognyan Kulev's patch for large stores, finally killing off the legendary 2 GB partition limit.

Upstream Hurd development continues to be very seasonal. Thomas and Roland have mostly disappeared again, but Marcus and Neal are now working on the L4 port with increased energy and the progress they make is awesome. First, Neal finished the initial memory management framework. A short while later, Marcus reported the first successful execution of a user-space program on Hurd/L4. Marco mostly focused on Grub2 over the last months but lately came back and finished the console repeaters and the Hurd part of DHCP support.

However, there have been a lot of new people getting involved both in Debian GNU/Hurd and the Hurd/L4 port, together with the Hurdfr project getting more active recently. A constant flux of incoming patches and some duplication of work resulting from this forced us to turn on the patch tracker of the pkg-hurd alioth project. Likewise, the hurd-l4 mailing list has seen a couple of new faces, most notably Matthieu Lemerre (who already sent in a bunch of useful patches) and Johan Rydberg (who has been active in the Hurd community for several years and now got commit rights to the Hurd/L4 source). Also, Andrew Resch managed to boot into the Hurd from a CD via GRUB and Ben Asselstine even got it booting all the way into the Hurd console and multi-user mode. This is an important step on the way to a native installer.

The Debian GNU/Hurd port is currently pretty well off, we have a stable toolchain and noticably increased the user experience lately as mentioned above. However, the sarge release is more and more becoming annoying for us, as important packages like coreutils or krb4/krb5 (which are needed for openldap and thus a lot of other stuff) are not getting fixed for us. Last week, I spent two nights fixing xfree86 for GNU/Hurd, but the patches did not get considered for the latest upload, which is pretty demotivating. Besides this, the next steps are probably looking at debootstrap and later debian-installer and then porting GNOME.

After having been demotivated to help on the sarge release due to the lack of progress on testing-security over summer, I got back into bug fixing in autumn when I had the feeling that the ftp-masters and buildd wizards were creeping back out of their holes and some slow progress was being made. I attended the Bug-Squashing Party in Frankfurt in late November, driving there together with Andreas Barth and meeting Frank Lichtenheld, Wesley Terpstra, Norbert Tretkowski and others. I took a look at RC bugs for the first time in 6 months and managed to fix quite a few. The connection between release progress and my bug-fixing motivation occured to me only afterwards, I wonder how many other Debian people were sub-conciously putting off fixing bugs due to being frustrated by the release process. The total lack of attendence at the spring Bug-Squashing-Parties on IRC suggest there are quite a few, so I was really happy to see people from the UK and Australia participate as well.

Next stop: FOSDEM.

1 Nov 2004 (updated 2 Nov 2004 at 10:12 UTC) »
Linux World Expo

After Systems, I spent a couple of days with my parents in Frankfurt and seized the opportunity to visit Linux World Expo last Tuesday and Wednesday. I was rather late on Tuesday and arrived at the Messe around 4:30 PM. The Debian booth was quite bigger than the one at Systems and professionally setup by Jörg Jaspert and Alexander Schmehl with our shiny merchandise display cylinder and a couple of impressive workstations (including one from Sun) as well as a small FAI part. The booth was manned by enough people, so I decided to have a stroll over the expo area, after greeting Chris Halls and Jesus Climent, who I was delighted to see again. IBM, Sun, HP and Novell (again, no GNOME desktops could be seen) had big booths, SAP and RedHat somewhat smaller ones.

I had a long and nice discussion with the two people from the C&L booth (the printing house which provided our booth at Systems), and they said that Systems seems to be interested in having a much bigger FLOSS area for next year, hopefully including the GNOME, Mozilla, Apache projects besides many others. This year, the FLOSS area was a bit hurried, but hopefully it will bigger, better and more shiny next year.

I also had some chats with a guy from IBM and another guy from Univention about commercial Debian support, which seems to be getting better and better in Germany. The Univention people seem to be able to power large installations. Of course, the IBM guy said they are not going to advertise Debian support, but the general rule these days seems to be: "If customers ask for Debian, we do support".

After the expo has ended we had dinner in an Indian restaurant. Michael Meskes, Jörg Jaspert Alexander Schmehl, Chris Halls, Thomas Lange and a couple of others were present and it the food was very good (albeit too much and quite expensive). Michael and Chris also took this dinner as an occasion to celebrate the start of their UK branch, Credativ ltd. Afterwards everybody left, so I decided to go home as well and had an interesting disussion about Ubuntu, Canonical and Debian with Chris while waiting for the train.

On Wednesday, I was unfortunately very late for Mark Shuttleworth's keynote. I only managed to listen to last five minutes and the Q&A part. However, probably knowing much of the content already, I wanted to attend the keynote mostly to see Mark speak anyway, which I was able to evaluate at least roughly during that time. Mark is a very charismatic speaker who seems to be well experienced in delivering talks. Also, the Q&A session was very enlightening by the way in which he answered, even though not so many questions were asked. After the talk, I had a chat with the present Credativ guys (Andreas Müller (amu), Michael Meskes and Chris Halls) and then went to lunch together with them and Mark in a nearby Pizzeria. Mark is really an interesting guy with a strong vision and nice humour and I can now understand why people would follow him and work with him so well.

Back at the expo, I bumped into Frauke Lehman, a socialogist from Berlin. She wanted to ask a couple of questions about the Debian project and as I was the first one around at the booth, we started a conversation. I think we talked for over an hour; about Debian history (mainly how the current organizational structure came into place), internal Debian mechanisms and communication channels and random other stuff. A very intersting conversation overall.

I needed to get back to Munich later that day, so I had to leave Linux World Expo at around 4:30 PM already. Still, it was a pretty intense event compared to one week of just hanging around the Debian booth at Systems, telling people what Debian is about. I had enough expos for a while now, though.

25 Oct 2004 (updated 2 Nov 2004 at 13:40 UTC) »
Debian at Systems 2004

After a two year hiatus, Debian (along other Free Software projects like Skolelinux, KDE, OpenOffice.org and the various BSDs) had a booth again at the Systems expo in Munich, thanks to the C&L publishing house who sponsored the Free Projects Area. The booth was operated by the Munich Debian crowd, namely Jens Schmalzing, Robert Lemmen, Erich Schubert, Michael Ablassmeier, Richard Atterer, Achim Bohnet, Simon Richter and myself. The booth itself was pretty small, only one half of a round table, about one square meter in total, but at least there was some wall space around where we could place one Ayo poster, along with another one right above the table. Unfortunately, our booth was located quite in the background and was not very easily visable from the main conference corridor, so over the days we added some stuff like a sign with a big, A4 sized 'debian' on it and a A0 poster version of the Debian flyer next to the one from Ayo (those Ayo posters are really cool, but they lack a bit in contrast, so they are hard to identify as Debian from far away). As promotion material for passing visitors, we handed out Debian flyers and LinuxTag DVDs (the latter for a voluntary donation, when we felt people would only throw it away eventually anyway) and we sold the above mentioned Ayo posters (quite some people requested T-Shirts as well, though). All the merchandise/marketing material was kindly provided/shipped by Credativ.

Despite NetBSD being around as well, the Debian booth had the coolest piece of hardware, namely a Mac SE/30, running Debian stable on a Linux-2.2 kernel. Jens Schmalzing got it installed and running over the last months, so we were finally able to showcase it. As a general-purpose demonstration/information machine, we gladly accepted a Shuttle XPC box with a pretty big LCD which Shuttle donated to all the free projects. Additionally, my ThinkPad was around most of the time. While the Mac was of course running text mode (debroster most of the time), the other boxes ran the Debian GNOME desktop. For some time, we also demonstrated Debian GNU/Hurd running the XFCE4 desktop on both the Shuttle and the ThinkPad.

Jens Schmalzing and I went to the expo area on Sunday and started building up the booth. However, the Shuttle boxes had not arrived yet, so this was pretty much limited to setting up the SE/30. and depositing the information material in the storage area behind the booth. Systems started for real on Monday and was pretty busy setting up our booth a bit more at the beginning. Some time later, Robert Lemmen arrived and we started to shift our attention to the visitors. Attendence was pretty low Monday morning though, so I started to worry Robert came along for naught. He then invented our guerrilla marketing campaign, moving our presence to the edge of the Free Projects Area, next to the conference corridor. He put a chunk of flyers and two empty DVD covers on the prospect case, stood next to it and started talking to passing expo visitors. This proved very successful, and we were quite busy for the rest of the day (or rather, week). Luckily, Erich Schubert arrived at around noon and had time to install Debian on the Shuttle which were dilivered by then. Once somebody wanted to know something specific or people found out about our booth by theirselves, one of us moved back to the booth. It turned out that if at least one visitor was standing at the booth, more people got interested and stopped by as well.

The rest of the week went pretty smooth, we managed to have at least two guys for almost all the time. There was a quite higher attendence and interest than we (or at least I) projected (and a much higher interest than for the booth right opposite to ours), so we were usually pretty busy answering one guy's answers while two others patiently waited for their turn. The posters sold very well and a lot of people donated some money in exchange for a DVD as well, so collected a good amount of money for Debian. Besides that, it was much fun to man the booth, albeit exhausting. Of course, by far the most frequently asked question was: 'When will Sarge be released?', followed by a small amount of 'what about amd64?' and 'wow, it exists?!?' questions when people saw Debian GNU/Hurd running. The rest of the questions were a variety of more specific ones. It was interesting to note that although Systems is a general-purpose computer expo (and looking around, almost all displayed computers ran Windows), except for very few visitors everybody knew about Linux and considerably more than half of the passing crowd knew about Debian at least from hearsay. Also, a lot of people (probably having read the announcement on the events mailing list or Debian Weekly News) were die-hard Debian users/admins who just came by to say 'good work' or talk about how they use/love/hack Debian. Also, Andreas Barth, Rene Engelhard and Norbert Tretkowski visited us briefly at the booth.

On Tuesday and Thursday, the Debian crowd got together after Systems in the Augustinerkeller right in the heart of Munich. It was quite crowded (eleven people in total) and the mood was really good on Tuesday, the first people arrived at around 7 PM and Achim and I finally left the pub at 11:30 PM after some beers. On Thursday, Jens Schmalzing, Siggi Langauf and me were around (technically, Simon Richter was around as well, but he was busy discussing things in another part of the pub) together with the KDE and Skolelinux guys.

Overall, the Systems was both much for fun and exhausting than I expected, and I hope we will get some new Debian users soon(and maybe even some new Debian GNU/Hurd users, who knows). Thanks to all the people helping with the booth, most notable the non-Debian Developers Robert Lemmen, Achim Bohnet and Michael Ablassmeier. And again a big thanks to C&L for donating the booth in the first place, as well as Credativ for the merchandise and Shuttle for lending their hardware.

15 Oct 2004 (updated 20 Oct 2004 at 20:21 UTC) »
The Ubuntu development model

So, everybody says Ubuntu rocks. I tested it a while ago (and my parents now use it) and I can confirm that. Enough praise has been told about the Ubuntu distribution, so thought I would rather write about the Ubuntu project a bit, as this appears to be much more blurry.

It seems Canonical managed to pull off with a tiny workforce what Debian was not able to do with a thousand volunteers. Of course, there is the mythical man month: about three dozen highly skilled and motivated developers working full time on Ubuntu can somewhat compensate for thousand volunteers of which only a tiny fraction care about releasing at all. However, Ubuntu also bravely decided to take new approaches to distribution development (at least compared to Debian) and try fundamentally different ideas, a couple of which were taken from how the GNOME community works.

The following are the key development point I as an interested outsider gathered from reading their website and from following their mailing lists and IRC channels (corrections/additions welcome):

  • They have an infrastructure similar to the Debian one. The archive tools seem to be the same, given that their prime authors work for Canonical and they advertise python as their language of choice. They also have build daemons at least for their three supported architectures. They use mailman for their lists and (for now) bugzilla as their BTS though.
  • They split the Debian distribtution into 'main' and 'universe', adding additional packages to the former. The 'main' part is their supported set of Free Software, which comprises their base system, their GNOME desktop and their selected servers and console programs. Universe is basically all other Debian packages which happened to build for them. Next to those, they introduced a 'restricted' component which contains firmware and binary-only drivers.
  • At least for their first release, they branched unstable around halfway through their release cycle and worked from there. They concentrate on stabilizing and integrating their main packages and only worry about universe when there is time or interest.
  • They rebuild the whole archive (including universe) on their build daemons. If needed, they upload new versions for the packages, sometimes many revisions.
  • They have a set of rules that says they should be respectful and communicative between each other. Disputes are regulated by their technical board and community council. This warrants a good working climate between the Ubuntu maintainers, which makes Ubuntu fun to work on.
  • There is a rigid, time-based release schedule. Not only the release date itself is fixed well in advance, but also every major milestone along the release process.
  • As the release draws near, their release managers have to approve all new upstream versions and later on every fix. They also produce several testing CDs in short time intervals.
  • Everybody involved can upload any package (as long as the patch gets approved), there is no concept of NMUs (Non-Maintainer-Uploads). More precisely, there is no concept of package maintenance at all, different developers are just loosely appointed to specific parts of the archive, like X, GNOME, etc. Furthermore, they have teams working together concentrating on specific tasks like server, laptop, security, installer, ports, etc.

Especially the last three points make their release management both more flexible and more rigid at the same time compared to Debian's, and they allow for their strict 6-month release cycle.

It remains to be seen how the Ubuntu development evolves. Some interesting questions in this regard, which will only be answered by time:

  • Will all of their work flow back upstream? They modified the standard GNOME desktop, the Debian installation and the project-utopia stack a fair bit, along with numerous smaller tweaks. Whether some or all of those ideas will be proposed for inclusion (and accepted) will be seen.
  • Will they seriously try to integrate their distribution-wide modifications into the Custom Debian Distributions framework, or will Ubuntu rather become a distribution framework for Custom Ubuntu Distributions themselves?
  • Will they branch off unstable again, or will they just resync the untouched packages? They seem to have modified a huge amount of packages and no central source-repository-management (yet?), so branching/merging might be too cumbersome for them and they might just work from their release.
  • Will the Canonical employes who are also Debian Developers continue to dedicate some parts of their free time to work on their non-packaging related tasks for Debian, like account/release/archive management?
My answer to all of the above questions was initially 'yes', and I hope this will continue to be the case. But again, only time can tell.

Of course, it also remains to be seen how Debian development evolves. Ubuntu seems to be the first evolutionary challenge to Debian and it will be interesting to see how Debian adapts to it. It is already clear that Ubuntu will be good for Free Software in general and the Linux desktop in particular, no matter what happens.

18 Sep 2004 (updated 18 Sep 2004 at 18:39 UTC) »

I am going to start my Ph.D. next month (well, in case I find a new flat in Munich until then), so I decided to do some Debian hacking while I still have some time. I finally managed to build and upload untampered hurd-i386 xfree86 packages for the first time in history, this should make it much easier to build other packages now. Jeff Bailey said he would restart the Hurd buildd in the near future, so the port should start rolling big time again soon. I also assembled and uploaded a new Hurd package, fixing the permissions problems invoked by the switch to cdbs and adding Marco Gerard's keyboard and mouse repeaters.

Besides that, I also announced a new Debian GNU/Hurd base tarball for xattr-hurd-enabled Linux kernels. Grab the kernel-image or use the kernel-patch, extract the tarball with star, adjust grub and boot into a functional GNU Hurd. It was never easier to get it than now!. The gory details are still here. Somewhat relatedly, Philip Charles has released the K7 series of the Debian GNU/Hurd ISOs.

On the OpenBabel front, I implemented Turbomol support some time ago. I would like to work more on OpenBabel, but Geoff apparently is going to merge the new conversion framework soon, which will result in some shaking. Otherwise, I also cannot get off my ass. I really think OpenBabel should extent to being a base for other chemical applications like ghemical or gchempaint. Those currently use OpenBabel only for import/export of files and then transfer the structure data to their own internal representation. Interaction between the different programs would be much easier if a common ground was used.

I've spent last week in Frankfurt. My parents weren't at home for the most part of it, so I had a pretty relaxing time with my notebook hooked up to our DSL/WLAN router sitting in front of the TV watching the olympic games or outside in our garden, while I abused my parent's P4 box as build machine. After I managed to build GNU Mach with Alfred's NIC-update, I was even able to SSH into it while it ran Debian GNU/Hurd. I finally managed to build glibc for hurd-i386, it took a ext2fs server compiled with a glibc from CVS to circumvent the nasty linkref bug. I also built new binutils and gcc-3.3/3.4 packages, getting the hurd-i386 toolchain back on track. It seems the only thing missing is an update of the hurd package itself, switching the kick-ass Hurd console on by default. Then, the K7 set should be able to go gold.

On Saturday, I passed by the real-life Debian Bug-Squashing-Party, meeting a couple of friends there, most notably Frank Lichtenheld and Peter de Shrijver. The party took place at the Lichtwiese in Darmstadt, the place I passed my first two years of studies in the last 90s. Well, I was not able (or perhaps motivated?) to squash bugs a lot, but just hanging out there during the barbecue was fun enough. Unfortunately, I forgot the AC adapter for my Thinkpad, so I was more or less forced to leave the party at around 2 AM. Many thanks to Martin Zobel-Helas and the others for organizing it.

I also finally did some non-Debian hacking again, adding Turbomole file format support to OpenBabel (CVS commit pending) and hacking on the libghemical autotools setup. It was nice to see that some guys from the University of Iowa are working on ghemical as well, they try to interface it better with the GAMESS software package. However, I believe that support for electron densities and vibrational frequencies should really be included in OpenBabel and hooked into ghemical from there.

22 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!