Dear Aunt Google, //at 23:31 //by abe

… this is “Do no evil”—urchin.js isn’t.

For luck, urchin.js and friends can be easily blocked using e.g. Firefox plugins like AdBlocker or NoScript, or with filtering proxies like Privoxy. And a line like

address=/google-analytics.com/0.0.0.0

in the dnsmasq.conf of your home router works like a charm, too.

SCNR, via Symlink

BarCamp Zurich — Resumé //at 02:02 //by abe

The BarCamp Zurich 2006 is over. On the way there I thought about what I would do during time slots with no interesting talks. But when I tried to make up my personal schedule, I noticed that I rather would have the opposite problem: Too many interesting talks at the same time… Well, to many interesting talks at all, although I only went to tech talks and left out the biz talks.

I first went to the Podcasting & Co. talk by Timo Hetzel, since I never heard or made a podcast, but was curious about podcasts in general. Besides statistics and rankings he spoke about where people listen to podcast (most listeners seem to do that during commuting), what people like in podcasts, why companies podcast, etc. And that a very big share of all podcast listeners use iTunes as podcast client and except juice (never heard of it before) all other podcast clients seem to be irrelevant.

My conclusion: I haven’t missed anything not having listened to or made podcasts neither do I need to listen or make podcasts in the future. They’re irrelevant. To me. :-)

Then I had to choose between the talks AJAX@localhost (PDF) by Harry Fuecks and Realtime Collaborative Text Editing and SubEthaEdit by the Coding Monkeys. I heard about realtime collaborative editing once know that it’s a challenging task for the developer. I also know what AJAX is (and that I would only use or recommend it for bells and whistles, but not for content in general), but “AJAX@localhost” sounded like writing normal applications using AJAX. It sounded interesting and evil at the same time. I had to go there! ;-) Others had similar expectations after reading the talk’s title, so I was quite surprised that it was about something completely different, namely about debugging AJAX on the localhost but under conditions usually only appearing if you’re running AJAX application not from localhost but from somewhere on the net: You may have different lags with every request, so some requests may reach the server before others, which may screw up the whole AJAX application, if the developers didn’t think about it and only tested it on localhost. (Hence the talk’s title…)

My conlusion: I will use and recommend AJAX even more seldom, since there seem to be even more design misconceptions than I thought before. But I’ll once have a look at the Webtuesday meeting, he mentioned.

For the third time-slot, I didn’t need long to decide where to go: I already knew a little bit about Microformats and I wanted to know more. Tag Trade also sounded interesting, but the second part of the talk’s title, Paid Learning sounded like business and so I had no scruples to cold-shoulder that talk. I probably didn’t learn anything really new in the microformats talk, but my knowledge about microformats is now more concrete, and after talking with Cédric Hüsler later during a break, I would even trust myself to start and define a new microformat.

Then I went to the HG Caféteria together with Gürkan and two German guys. While waiting in the queue, we were talking about our jobs and our favourite Linux distributions. I got some rhubarb pie and a rum truffles, assuming that the Caféteria uses no alcohol in their products like all other SV restaurant I know. But this one seemed to have quite a lot of alcohol, since it felt like my breath was burning… Well, this resulted in my second SV feedback form submission…

Next I went to Alex Schröder’s talk about multilingual websites, Oddmuse and the Emacs Wiki, although also the talk A-Life about simulating evolution sounded promising. Alex asked the listeners about their experiences with multilingual websites and showed what Oddmuse offers as partial solution to the general multilingualism problems. But regarding the comments from the auditorium, there probably won’t be a perfect solution until computers can translate perfectly…

The next talk I visited was Gabor’s talk about his master thesis Organizing E-Mail which resulted in a soon to be released Mozilla Thunderbird extension called BuzzTrack. From the other concepts he showed, I found Microsoft’s SNARF (Social Network and Relationship Finder) and IBM’s Thread Arcs most interesting as well as the fact that there is no e-mail client seems to have a majority at all.

Directly after Gabor I had my own talk about Understanding Shell Quoting, so I also couldn’t go to Adrian Heydecker’s talk about Learning with Hypertext and Search Engines. I had only about three and a half listeners of whom several to my surprise where here because they didn’t know what “shell quoting” is.

I really didn’t expect that.

But that seems to be one of the differences between a BarCamp and a Linux Conferences: People come here to see something new, something they haven’t heard about before. On Linux events most people come, because they already heard about some special topic and want to know more or learn something about it. On Linux event my shell talks usually were attracting many visitors while at a BarCamp, talks presenting an idea, a concept or a tool seem to much more interesting for the attendees. So for the next BarCamp I perhaps exhume my Website Meta Language talk which never seemed to hit the nerve of Linux event attendees, since it tried to “sell” a different concept of generating website than most were used to.

At least one listener excepted the talk to be named “shell escaping”, but IMHO escaping is only one quoting technic and it’s not only used for quoting. But perhaps I should take the word “escaping” in the title though for the next time.

Happily most of the listeners seem to have learned something new from the talk and Silvan Gebhardt was really happy about his new knowledge about ssh ~ escapes, although I mainly talked about how to quote them than how to use them. :-)

During the last slot I visited the session about the upcoming BarCamp Alsace 2 and the yet to be planned BarCamp Rhine, a BarCamp to be held on a ship traveling from Basel in Switzerland down the Rhine, stopping in Strasbourg, Karlsruhe, Rhein-Main-Area and perhaps even Cologne and Amsterdam.

Contrary to my initial thoughts, the day was over very fast and I had no single boring minute during the BarCamp. Wow!

After we’ve been kicked out of the building by ETH janitors, we joined again at the Bar N-68. On the way there I met Urban Müller who attended BarCamp Zurich, too. We talked quite a lot and it was very interesting to see behind the scenes of e.g. map.search.ch. Later I joined the French speaking table, talking with Gregoire Japiot from WineCamp France and Alex Schröder.

Around 9pm I left the N-68 as one of the last BarCampers, tired but with new knowledge, new ideas, new acquaintances and a new hobby: BarCamping. What a luck that BarCamps aren’t that often, otherwise I couldn’t afford this new hobby. ;-)

As a relaxing end I met with Alex Schröder and Christophe Ducamp on Sunday morning for brunch in the restaurant Gloria in the Industriequartier. When we were leaving the Gloria I noticed their book board with a lots of BookCrossing books and I took “The Da Vinci Code” with me, since I saw the movie and people were telling me that the book is much better. I’ll see…

Yet another old laptop //at 04:13 //by abe

My father got me a nice IBM ThinkPad from 1996 earlier this year, so the next old laptop he digged up was planned to become a christmas present for my brother. But my father didn’t manage to find out, how old nor how fast that laptop was. And when I found out that it was a Pentium I with 90 MHz, it was clear, that my brother wouldn’t have any use for it, so he got “only” the used 850 MHz AMD Duron midi tower and my parents declared that old Compaq LTE 5100 laptop as a christmas present for me. :-)

As my IBM ThinkPad bijou, this Compaq LTE 5100 is from 1996 and has a Pentium I processor. Both also have a 800×600 resolution, a double PCMCIA slot and a floppy drive, which can be replaced by a CD-ROM drive (if I had one). But that are all similarities. Technically the Compaq has 90 MHz instead of the ThinkPad’s 133 MHz, but therefore has 72 MB RAM in comparison to the 48 Megs the ThinkPad has. Also regarding disk space the Compaq outperforms the ThinkPad: 1.6 Gigs of disk space in comparison to the ThinkPad 1.0 GB hard disk. Another difference is the battery: While the ThinkPad can work over 2.5 hours without external power, the Compaq even didn’t manage to completely boot its currently installed Windows 98 (the ThinkPad had a Windows NT installed when I got it) when running on battery. (Will do that test again when I can confirm, that the battery was full before testing. :-) Yet another difference is the keyboard layout: The ThinkPad has an US layout while the Compaq has a Swiss-German layout. But the most obvious difference is the look: The black ThinkPad still looks like having a modern design while the Compaq looks very very outdated in its perfect computer beige and with its quite small display.

So retroperspectively, it was a good a idea to name the ThinkPad “bijou” (French for jewel, jewellery, gem, etc.; named after a very neat british two-door limousine built in the UK by Slough on a 2CV base during the ’50s). Because now I have the choice between a lot of not so nice looking (not to say ugly ;-) 2CV derivatives to name the Compaq after. My favourites currently are the Iranian “Baby Brousse”, the Greek “Namco Pony” and the German “Fiberfab Sherpa”, all canvas and flatbed style 2CV based buggies, similar to the original Citroën Méhari but with steel body instead of the Méhari’s controversial plastic body. And one of the not used names, I can use for further ugly Compaq laptops¹.

Another question yet to answer is the question of what operating system to install on it. Since the ThinkPad runs fine with Debian 3.0 Woody and I have a lot of other Debian boxes at home (running Woody, Sarge or Sid), I currently think about installing the very fresh NetBSD 3.0 (released on Christmas’ Eve 2005), FreeBSD 6.0 (released early November 2005), DragonFly BSD 1.4 (to be released in December :-) or DeLi Linux 0.7 pre (which was also released in early December 2005 and already uses X11R7). Another idea was to install grml 0.5, but since grml is a live CD distribution, it probably would be hard to install it over network. Same counts for ReactOS (version 0.2.9 was released shortly before Christmas 2005), which doesn’t seem to have a floppy disk plus network install. Since I always planed to upgrade my currently defective Toshiba T6400 i486 laptop ayca (maybe after getting an organ donor on eBay or so) to DeLi Linux 0.7 (and perhaps write a review about it for Linux Magazine or so) and I may get an Sun Ultra Enterprise 2 soon (on which NetBSD 3.0 would be the perfect OS since Linux’ performance still seems to suck on Sparc :-), I currently prefer the FreeBSD or DragonFly idea. If the Ultra doesn’t come, it probably will get NetBSD, since I haven’t a NetBSD box yet. (Haven’t a DragonFly box either, but a FreeBSD 4.x running somewhere. :-)

Well, I guess, I’ll take even more old laptops than last year to the Vintage Computer Festival Europe (VCFe) in Munich next May. And since the two 1996 laptops are now 10 years old, they’re even ontopic! Yeah! ;-)

¹: I have two other not yet working Compaq laptops, both from an elder generation than Pentium I. One I got on a Swiss flea market for a few euros and the other was the first laptop of my boss, which he else would have thrown away. Unfortunately both are without power adapter and neither the usual allround laptop power adapters from Conrad, etc. nor the one from the LTE 5100 fits. But since there is eBay, I expect to get such a power adapter once. :-)

X on IBM ThinkPad 760ED //at 02:39 //by abe

As many of my friends know, I installed Debian GNU/Linux 3.0 aka Woody on my IBM ThinkPad 760ED (Pentium 1, 133 MHz, 48 MB RAM, 1 GB HD) named bijou at Chemnitzer Linux-Tage this spring. Although many helped me trying to configure X, I didn’t get rid of the LCD “forgetting” more or less pixels each row so that the picture got blurred towards the right rim.

During Berlinux, a not so convinced Linux user (he told me he likes the feeling of Windows, but got stick with Debian because his modem just didn’t work with Windows) told me that he had similar problems with a graphics card with similar chips as the ones in my ThinkPad. Since I never was sure, if my problem is a hardware defect, a driver or a configuration problem (but I tended to hardware defect or a not supported chipset, since I read about several Trident 96xx and 98xx chipsets to be unsupported under Linux), his comment was a ray of hope to me. He told me he found the solution on Werner Heuser’s TuxMobil website, whom I showed my X problem also Berlinux but who hadn’t an idea what it could be. He also told me, that XFree86 4.x doesn’t support this graphics chip, but XFree86 3.3.x does.

But somehow I had forgotten that TuxMobil not only has informations about Linux on laptops but also a big bunch of links to pages which deal with specific models. I can’t remember, if I looked there already back in March, but it felt like I didn’t although my own small text about bijou is linked there, too. I looked through the other 760/770 ED/XD pages and on the second or third I found someone who seems to have had the same problem and also with a 760ED. He wrote, someone else has gone down that path already so he linked directly to the XF86Config he found elsewhere. That sounded like an easy earned money so I followed the link — 404. Shit! For luck a few lines down he linked also his own XF86Config, so I grabbed it, uncommented everything unnecessary and put in the essential parts of his XF86Config of which the most important parts probably were the modelines. The one for 1024×768 was commented as the only one working, those for 800×600 and 640×480 were commented out. Then I downgraded X to XFree86 3.3.6.

It didn’t work as expected. The display stayed black which was less than I had accomplished before. Shit! But giving up is not my style. So first I reduced the color depth. No change. Then I started with reducing the resolution. With 800×600 and 16 bit color depth, it finally worked. No hardware defect, no unsupported graphics chip. Just not the right modelines. That was all. YESSS!

I guess the guy whose XF86Config I used didn’t have a 760ED but a similar model with a LCD with higher resolution. because my 760ED definitely has no 1024×768 resolution because 800×600 fills the screen completely.

Still leaves the problem with the svgalib: The system just freezes with a black screen if I start a svgalib application like e.g. zgv. First I found out, that svgalib indeed has a configuration file which (at least under Woody) can be found at /etc/vga/libvga.conf. Copied the modelines from the now working XF86Config, configured the mouse and tried again. Freeze. Hmmm, in the config there is mentioned that if svgalib doesn’t correctly recognise the graphics card’s chipset, you can hardcode it with a configuration directive, e.g. “chipset VGA” for otherwise unsupported chipsets. And that worked, although only with 640×480 yet. So I tried the only setting for Trident cards found in the list of supported chipsets. And what happend? Right, the system froze again. So svgalib probably recognised the card as Trident and used the only available Trident driver which was obviously the wrong one. So here are my XF86Config and my libvga.config working on the IBM ThinkPad 760ED.

But nevertheless — this 0€ laptop has just proven that it can be even more useful than it already was with text mode only. I also already played Frozen Bubble up to level 25 or so on the train back from Berlin. Old hardware rules.

But I now also have another problem (again): Since X works now, I can run Galeon 1.2 on the ThinkPad, but GTK 2 respective GNOME 2 are much slower than in the 1.x versions and also need much more ressources, which the laptop just does not have. And since I — as most of the people who read my blog or Planet Debian should know ;-) — don’t like Galeon 1.3, I probably won’t dist-upgrade my ThinkPad to Sarge that fast although I already thought about it. XFree86 3.x isn’t in Sarge either IIRC but this should be no problem since the Woody packages are said to work under Sarge, too. Well, still yet another reason for forwardports.org… ;-) Or maybe I can get Kazehakase running on Woody so I can drop at least the whole ballast GNOME 2 comes with. We’ll see…

JFTR (Update on 2nd of May 2007): My current XF86Config-4 for Sarge on my ThinkPad 760ED named bijou.