Continuing on the reading binge, I’ve read: The Warren Buffet Way by Robert G. Hagstrom: This book got really repetitive after a while. Warren Buffet is a disciple of Benjamin Graham, the original value investor, who is referred to heavily in all three books. According to Graham, the theories of an efficient market don’t apply very well, and it’s occasionally very easy (especially during a recession) to buy companies (or their stocks) for far less than they’re worth. Buffet simply buys the whole company. The best advice is that “the market is only useful as a reference to see if anyone’s offering to do something foolish”. It’s a theme that’s echoed in all three books. One Up On Wall Street by Peter Lynch and John Rothchild: Not as good as I’d hoped, but in some ways better, and worth reading. The title comes from the idea that individual investors are actually better placed to take advantage of the market before institutions, since they often have access to information to assess a company with before analysts even take a look at the company. He then follows this up by saying that you should treat it as a tip, and that you should treat all tips the same: as a reason for investigating a particular company. All of the preceeding didn’t impress me much. But he then tells you how he goes about assessing a company, especially reasons why he won’t buy a company, which I found very helpful. The Intelligent Asset Allocator by William J. Bernstein: The best of the three. This book picks up where Random Walk leaves off; that book encourages diversification, while this one shows you how to do it. Also, Random Walk focuses largely on the stock market, whereas this book discusses lots of asset classes and the correlations between them and how to pick a portfolio designed to maximize return while reducing risk. Also he spends a lot of time on specific details about which mutual funds are good for various asset classes, as well as various tips for really squeezing out as much return as possible through thinks like tax-conscious funds.

Security Updates

A new version of GNUMP3d is out. This is primarily a security update. Install it now if you use it publically. (Local mirror in case you’re running from source and the gnump3d.org site hasn’t updated yet)

Apt Updates

I’ve updated my personal APT repository. This now uses reprepro to create/update/manage it. The reprepro tool is utterly wonderful. (Although I used the backported Sarge version, rather than the older one). Expect an article on the Debian Administration website shortly. Now I have an “incoming” directory I can upload new package builds to, and have them magically migrate into the relevent section. The other significant change is that now I can host packages with Arch: All, as well as i386. This is useful as I distribute several perl scripts/libraries which were previously unavailable to non-x86 users. Talking of which I still have to finish my review of The Debian System written by Martin Kraft - I’ve been sitting on that for too long. (Summary: Great Book, buy it if you can’t get one donated to review ).

board games Bought myself the European version of "Ticket to ride" in preparation for the winter to come. Also, some Carcassonne updates ("The dragon and the princess" and "The count") since i believe i have all the rest. Maybe i should also get "The city", but it is a standalone game, so i need to read some reports. dspam Dspam is back. We have a coordinated effort to get it out in less than a week. I had a package more or less ready, but since two non-developers are working on it, too, i stepped out and right now i am more an observer than an active packager, while i clean (read janitor) some things from time to time. Also, too much work is keeping me from doing other Debian stuff, including coordination of DC6. music I have recently discovered beatport.com. I have refused to access it before, due to its Flash (yep, chilling effect) interface. It is crack for the following reasons:

• One cannot set bookmarks. I would like to set a flag on some artists, or just have a "duh, this one i have to recheck". Forget about it.
• One cannot open multiple windows/tabs. While I am surfing the interface, i would like to open some labels/artists in another tab for inmediate review. CAN NOT DO.
• Shit, man, one cannot even send a URL to a friend for him to buy a song, since there is no such concept in the interface.
• Also, one cannot go back, since the flash buttons (RW, FF) are b0rk3n.

So, please, beatport, if you are out there and listen to your users, do something about it. However, the music selection they have rocks (even if it is crappy electronic music).

Here's how to subscribe to mailing lists with a combined total posts of 2000 or more per day, and live. It's all about pattern recognition. Threaded mail readers such as mutt sort emails into a tree based on which emails are replies to other emails. The mail index displays this tree of messages with the author of each indicated. Since the Subject field rarely changes, mutt does not display it unless the thread's subject changes. Combining these elements of author names, thread trees, and occasional subject changes produces patterns that the brain can use to quickly identify important and useless sets of mail messages, without actually reading anything, and without resorting to complicated and fragile killfile and scoring systems. Some patterns are very simple. For example, this is the "take it to private email" pattern. This pattern can be thread-killed on sight; if you're feeling generous, read the last two messages in the thread.

bob    +-> Foo
fred     +->
bob        +->
fred         +->
bob            +->
fred             +->
bob                +->
fred                 +->
bob                    +->
fred                     +->

This is the "think before you post" pattern.

bob    +-> Foo
bob      +->

This is the "blindingly obvious answer" pattern. If you're bored you can read any one message to get the gist. Sometimes many of the replies will be worth reading to get a broad overview of all possible responses to a given obvious question. Most times not.

bob    +-> Foo?
fred     +->
alice    +->
george   +->
nancy    +->
steven   +->
nick     +->

Here's a deceptively simple pattern, the "uninteresting message". It might be worth checking the post date to make sure that it's not just the beginning of one of the above patterns of course. And if you're feeling generous, you can read some small fraction of them.

bob    +-> Foo?

This might be a worthwhile thread. Notably because it's short, has a reasonable mix of participants, and seems to come to some kind of conclusion.

bob    +-> Foo
fred     +->
alice      +->
bob          +->
george         +->
bob              +->
nancy    +->
fred       +->

And here's a transformation of that same thread featuring "too late for the party syndrome"

bob    +-> Foo
fred     +->
alice      +->
bill       +->
bob          +->
george         +->
bill             +->
bill         +->
nancy    +->
fred       +->
bill         +->
bill       +->
bill     +->

This is a perfect time to er, threadkill bill. Or if feeling generous, read the second or third of his many replies to see what's got him so worked up. To see more sophisticated patterns, you need to categorise participants of the mailing list. A few people almost always manage to turn even a useless thread into something interesting, or have viewpoints that are good foils for your own. Learn to recognise the names of these interesting posters without thinking about it (I'll name them in uppercase below). Similarly, some people never contribute anything useful, and can be filed under the label of k00k. If there are more k00ks than interesting posters it's probably time to leave the list. Most people won't fit in either category. Some people change category depending on the subject. Here's the same potentially worthwhile thread as before, but now the people matter.

bob    +-> Foo
fred     +->
k00k       +->
bob          +->
george         +->
bob              +->
NANCY    +->
fred       +->

Suddenly, three messages of the thread become worth reading, and the larger k00k-tainted subthread's likelyhood of being valuable drops substantially. Similarly, this thread can be ignored.

k00k   +-> Foo
bob      +->
fred       +->
k00k         +->
alice  +->
k00k     +->

And this thread has a high potential for being interesting.

bob    +-> Foo
fred     +->
bob        +->
NORRIS       +->
george         +->
bob              +->
NANCY    +->
fred       +->

This thread, a variation of "take it to private email", increases bob's chances of being a k00k. Or he's really right and everyone else is wrong.

alice  +-> Foo
fred     +->
bob        +->
nancy        +->
bob            +->
alice            +->
bob                +->
fred                 +->
bob                    +->
fred                     +->
bob                        +->

This thread is a hard case to call, it depends on how good the interesting poster is in deflecting k00kitude. There could be as many as three worthwhile messages in here; I'll flag them.

  k00k   +-> Foo
  bob      +->
* NANCY      +->
  k00k         +->
* bob          +->
* NANCY          +->

It might be worth reading the interesting poster here, but probably only for a laugh, and it's just as likely that they've been cruelly trolled into answering, especially if an instance of the "blindingly obvious answer" pattern is grafted onto one of the subthreads.

k00k   +-> Foo
NORRIS   +->
k00k       +->
NORRIS       +->

So far the Subject has been ignored, and generally it's more important to look at the thread structure than the subject, but there are useful exceptions. It might be worth checking out the differently named subthread here, on the other hand its close proximity to the k00k likely means its a useless tangent.

bob    +-> Foo
fred     +->
k00k       +->
bob          +->
george         +-> Bar (was: Foo)
bob              +->
fred               +->
fred           +->
tony     +->
fred       +->

Here is not just one thread, but a collection of threads. This pattern, typically associated with a much larger thread than shown here, is a sure sign that a thread has gone pear-shaped and unless the followup thread has a lot of interesting posters in it, there's not much hope for the lot. (k00ks also spawn disconnected threadlets; such threadlets should be ignored.)

 k00k   +-> Foo
 bob      +->
 NANCY      +->
 k00k         +->
 bob          +->
 NANCY          +->
 bob    +-> About foo..
 tony     +->
 k00k       +->
 alice  +-> That foo thing

That's enough patterns to get started with, but they really come into their own when picked out of larger threads in an overloy long ongoing discussion. A few examples can be found in this segment of 23 lines of a much longer thread.

bob                                                       
k00k                                                      
tony                                                      
k00k                                                      
tony                                                      
k00k                                                    +-
bob                                                  
k00k                                                +-> Bar
john                                    +-> Bar
nick                                  +-> Bar
barney                                  +->
nick                                      +->
nancy                                       +->
NORRIS                                        +->
nancy                                           +->
SAM                                           +->
nick                                            +->
NORRIS                                            +->
oliver                                          +->
john                                    +->
frank                                     +->
john                                        +->
tom                                           +-> Mumble
bob                                             +->

Part of this thread is useless by two criteria; it has far too much k00k in it, and it's nested so deep that mutt doesn't bother to try to draw the thread indicators. The next bit though shows some signs of being interesting despite its fairly deep nesting. And the tangent might be worth keeping an eye on. Those are the thread patterns that I know I know. I suspect there are many others that I also know, since I can typically read mail with the d key as fast as mutt will scroll, stop at the interesting spots, and not miss much. If you know of other patterns, send them my way. Finally, even if you prefer to read every post, please do us all a favour, imagine your reply is already in the thread structure, and look for patterns like "take it to private email", "blindingly obvious answer", "too late for the party syndrome", and k00k influenced posts.

Steve wrote in his blog:

I’ve seen this argument before “Buy distribution of GNU/Linux and support free software programmers”. The only problem I have with it is that it is incorrect. Buying GNU/Linux distributions helps the vendors who created it, certainly, and may indirectly help pay for some free software in the sense that the vendors might ship free software they wrote (e.g. SuSEs Yast 2]). However plonking down real cash-money for a boxed set of SuSE gives no money to the people who created MySQL, no money to the people who created Firefox, no money to the people who created Emacs, Vim, Bash, and Catan/Pioneers, etc.

I think, in general you’re right. And if you — as you did :-) — take SuSE, it usually works. And you’re probably also right for most people who just know the big, commercial distributions. But what if you take a free distribution like Debian or some of the BSDs, e.g. OpenBSD? How much truth is in there then? Especially in the case of OpenBSD your view doesn’t seem work, because if you buy an (official) OpenBSD box, you pay the developers — or at least a few of them — of the operating system core and some mission-critical applications. But what if you take community based distributions like Debian? You distinguished between distributor and authors of free software. In my eyes especially Debian, but also some other community based distributions are both at same time. So IMHO you can put them on the author side of your view. And since many Debian vendors (at least those I saw) donate a part of the profit they make from selling Debian CDs or DVD to the Debian Project. Or they offer additional shopping cart items “Donation to Debian” if you order a Debian item. (Example: LinISO.de) Another question in this context would be, how the FOSS world would look like if there are or were no commercial distributors. It probably would be much smaller because some marketing and some lobbying would be missing. Although that’s the only implication which comes to my mind, I’m sure, there are many more possible views on this subject. But as I said, IMHO you’re right for most cases.

PJ@groklaw on SuSE 10.0:

You can download for the cost of your blank CDs, or you can purchase it for $59.95 if you are in the US, and you can buy online. I will probably do both, after reading the two reviews. I can’t wait, so I’ll download it, and then I will buy also, so that I can support wonderful FOSS software and because you get extras, like documentation in a book, which is good if you ever can’t get online.

I’ve seen this argument before “Buy distribution of GNU/Linux and support free software programmers”. The only problem I have with it is that it is incorrect. Buying GNU/Linux distributions helps the vendors who created it, certainly, and may indirectly help pay for some free software in the sense that the vendors might ship free software they wrote (e.g. SuSE’s Yast 2]). However plonking down real cash-money for a boxed set of SuSE gives no money to the people who created MySQL, no money to the people who created Firefox, no money to the people who created Emacs, Vim, Bash, and Catan/Pioneers, etc. So supporting free software distributors? Great! But to reward actual free software programmers? Go buy them a wishlist item, give them some cash, throw them a postcard, send them an email… Or not. As the case may be.

I spent a few days in Krak w, Poland for the CALIBRE workshop on Quality, Security and Safety and for KKIO 2005 and I used this opportunity to do some sightseeing. After the workshop on Tuesday, Matthijs den Besten and I wandered around the city for a a few hours, then went back to the hotel for a swim, just in time to go for dinner with the CALIBRE people. We went to a restaurant which served traditional food from the 17th century. It was a really amazing place with animal fur and carpets on the wall, big wooden tables and menus which had nice drawings on them. We had a selection of different food which we shared and it was really great. The following day, as part of KKIO, dinner was served in the Wieliczka salt mine after we went on a tour of it. It's pretty amazing what kind of art you can make out of salt. On Thursday, I did some more sightseeing, basically walking around the city center and looking at some of the nice attractions. Which turns out to be churches, given that this is Poland. I particularly liked the basilica of the holy trinity but many other sights, such as Wawel castle, were pretty impressive too.

I was at Berlinux 2005 this weekend and though the very chaotic — because understaffed — organisation it was interesting and also funny. Thursday I arrived around 20:15 in Berlin, met Klaus Knopper and others at the train station, headed to Sven Guckes’ appartment for dropping all my luggage, then going back to meet with Klaus and the others for a theremin concert with Dorit Chrysler. No wonder that it sounded sometimes like one of my favourite musicians, Jean Michel Jarre, since — according to the Wikipedia theremin article — he also plays this instrument. On Friday I held my talk about WML in front of a — for that topic — surprisingly high number of auditors (around 30, maybe 35). In comparision to my WML talk at OscomTag 2005 all people who asked questions had understood about what the talk was, so the questions were most time interesting and justified. As usual I held the talk using Lynx with LSS support (picture by Sven Guckes) on my nine year old Pentium 1 ThinkPad bijou running Debian GNU/Linux 3.0 aka Woody. Before and after the talk I helped out at Werner Heuser’s xtops booth (another picture by Sven) and the booth of the Debian Project (yet another picture by Sven :-) directly beside Frank Ronneburg’s Debian powered model railway. (picture by you-know-who ;-) In the evening I was at the social event, hanging around with alphascorpii, Tolimar and Joey and being surprised that Joey studies biology — as I did as minor to computer science. On Saturday I was on alphascorpii’s talk about why being a BOFH is not funny, hung around at the same booths as the day before, fixed the X configuration on my laptop after hints on a unknown Debian booth visitor. Before the exhibition closed I heard a very interesting talk about web accessibility held by Sebastian who is blind himself. Although or maybe because I’m interested in that subject, the talk opened my eyes regarding two things: First Captchas are evil and Blind HTML tables aren’t as evil as all the priests of web accessibility are always preaching . They are easier than frames for blinds and seem to have only little disadvantages against a CSS based layout for blinds nowadays if used the right way. Oh, and btw. — nested tables are still evil. :-) Saturday evening I had dinner together with Stefan Gerdelbracht, Frank Hofmann, Klaus Knopper and Manfred Krejcik. Later Thomas Winde joined us. It was very interesting evening, especially talking with Klaus and Manfred. On Sunday, after having brunch with Stefan and Manfred, we met with Sven (who was our host at Berlin, thanks again!) and shortly after that, Stefan left for visiting some other friends in Berlin. Sven, Manfred and I visited C-Base where Sven stumbled over a sound editing seminar while Manfred was preparing his zipFM show for Monday which mainly consisted of an interview with Klaus. After that we headed to a small but fine birthday party of a friend of Sven and were back home around 2:30. My train left Monday morning at 8:56 and I was at home around 14:30. And on Friday I’ll go to Dresden for the Linux-Info-Tag by train just to go back to Berlin afterwards, where I meet my parents for a two week baltic sea holiday in Mecklenburg-Western Pomerania near Rügen. But due to the Systems fair at Munich and autumn holidays I have to stay at work this week. And yes, I wrote this and the other postings posted today offline, so they’re dated quite close together. :-)

In Oldenburg, Joey gave me a PARISC box, an Apollo 715/50. It wasn't complete—no RAM, no hard disk, no CD-ROM or other media—but the system itself seemed complete and functional. Unfortunately, it isn't. I added some RAM that I took from my parent's old computer (196MB in total, IIRC) and tried power on the box. Didn't work. When I powered it up, the machine powers up its fan for a short while before shutting it down again; then the fan moves to a lower rate, and the machine appears dead. Well, mostly so; there's no output on the VGA port, and the LEDs don't do anything anymore, either. But those LEDs would appear to be helpful, if one would know what they mean. Unfortunately, the box came without manual, so figuring out exactly what they mean all by myself is going to be painful. Google to the rescue. I found an archive of the comp.sys.apollo newsgroup, with loads of postings from 2000. One of them also mentions failure to boot, no output, and mentions that all LEDs except one are on at the end. That's the same as what happens here, except that in my case the one-but-last one are on, rather than the last one. A reply suggests that the machine couldn't find the boot disk, and that therefore the box can't be booted. So, lemme get this straight: if the boot disk is having issues (or not available, as in my case), Apollo machines simply decide to not do anything? Fuck. Update: okay, turns out the HP Website still has a PDF of the model 715 service manual, which contains an explanation of the LED error messages. Turns out it means "No Memory Found". Hm. Watch this space.

The cast of thousands tasked with relocating us to the US continue to swing into action. For up to a month, we'll be accommodated at The Carlyle in Santa Clara. It looks a bit over the top. I mean we really don't need chandeliers, but at least it has a gym so I can try and keep off my Google pounds. I also looks reasonably close to work. Now if we can just get through all this customs paperwork for importing our personal effects...

Here's a bunch of smaller updates which probably don't warrant extra blog posts, but might be interesting nevertheless:

• On the right-hand side of my homepage you now have three new blocks which should simplify site navigation. Each contains a weighted list of tags (tag clouds) for my blog, my podcast, and my photoblog, powered by Drupal's tagadelic module. Of course, the "traditional" navigation menu is still there.
• There's also an RSS feed for comments on this site now, courtesy to Drupal's Comment RSS module.
• On the lower left-hand side of my homepage, there's another new block. It shows where the visitors of my site come from in (almost) realtime. This is a service provided by MapStats (I already talked about it earlier).
• I have uploaded updated versions of my bashrc, vimrc, and muttrc config files.
• My podcast is now listed in iTunes (thanks Daniel Reutter for adding it). You can subscribe by clicking on this button: . The amount of people who subscribe to my podcast have almost doubled, figures went up from 30 subscribers to more than 50 today! If you're reading this and subscribed through iTunes: thanks ;)
• My video iPod will (probably) be shipped on October 31, and it'll take a few more days until it arrives. Arghh, /me wants it now!

Mike Sperber finally reviewed my patches; they unfortunatly were not deemed appropriate for the Scheme48 distribution, see this thread. Hopefully these issues can be sorted out in the not-to-far future. If you also need the one or other features from my hacks so far, and are adventurous, you can conact me on #scheme on freenode for help with getting and building my fork. Writing portable Scheme programs with scmxlate and spells For the conjure project, we are using scmxlate to make it portable across multiple Schemes. For that purpose, I slightly extended scmxlate and built a kind of meta-module system on top of it. Here is how that looks like: You write a description of the modules your application or library is made of; this is done in a sightly extended subset of the Scheme48 configuration language. For example (from this page, see there for more details):

 (define-structure conjure.foo.bar (export qux baz
					  ((froboz fribnitz)
					  :syntax))
  (open scheme srfi-1 srfi-13 conjure.util.files)
  (dialect (guile (open ice-9.syncase))
           (scheme48 (open extended-ports)))
  (files (foo basic-bar)
         (foo bar)))

This defines the module conjure.foo.bar, which exports the identifiers qux and baz, as well as the macros froboz and fribnitz. The open clause means: this module is written in R5RS Scheme (scheme), plus additionally using srfi-1, srfi-13 and the conjure.util.files module. The dialect clause can be used for specifiying additional modules needed for the dialect-specific code for the module. Finally, the files clause tells us that this module consists of the files foo/basic-bar.scm and foo/bar.scm. This file then gets processed by scmxlate, generating a file with the right module definitions for the selected target dialect, e.g.:

(define-module (conjure foo bar)
  #:use-module (srfi srfi-1)
  #:use-module (srfi srfi-13)
  #:use-module (conjure util files)
  #:use-module (ice-9 syncase)
	
   #:export (qux baz)
  #:export-syntax (froboz fribnitz))
	
 ;; code for foo/basic-bar.scm and foo/bar.scm follows

will be generated for Guile. We recently split scmxlate and the meta-module systeem out of conjure, see the scmxlate category in the conjure archive, so it is available for other projects. Now the split-out of the general purpose modules into a separate project — SPELLS, (S)pell’s a (P)ortability (E)nvironment (L)ibrary for (L)ots of (S)chemes — is in preparation. This library will further aid writing portable code, by providing a consistent interface to functionality that has different APIs in different Schemes. These modules are already there:

• spells.util.pregexp: Dorai Sitaram’s portable regular expressions
• spells.util.named-args: Conviniently create procedures with named arguments
• spells.util.table: Hash tables
• spells.util.condition: SRFI 34 and SRFI 35
• spells.system.file: Lots of file-related proedures, e.g. copy-file, directory?, …

All of those are implemented for all of mzscheme, Guile, Scheme48 and Gauche, using native functionality where available. Other modules will surely follow, a project has been requested at Gna! to facilitate contribution. The spells library will also be used by my upcoming port of G-Wrap to Scheme48.

I’m going to get paid for open source developement! Wooha! More specifically, the company I work for since 2 months is going to switch away from CVS as SCM tool. We had a look several SCM systems, among them Subversion, GNU Arch and several commercial offerings, including BitKeeper :-0. The final contestants were GNU Arch and AccuRev. Subversion fell flat for its missing features (mediocre merging support, no distribution), and BitKeeper was not considered seriously because they don’t even list prices at their site (which indicates a fairly high price tag, normally and due to opposition from myself (for the obvious political reasons). Well, finally the case was made for GNU Arch, and since it lacks a decent GUI (currently the GUIish guys here use WinCVS), it was decided to develop the GUI in-house, and as an open source project!. It will probably be developed using Python and GTK+, so it will run also on Windows (which is a requirement). Expect more news, including project announcement soonish.

Congratulations to my brother, Casey, on passing his driving test yesterday (his first attempt, too, I believe... OK it took him 2 to do the theory, but hey, no-one's perfect ;).

So yesterday Google announced their earnings, beating estimates, sending their stock even higher. Every time someone says GOOG can’t possibly go any higher, they’re promptly proved wrong. So I’m not going to say that GOOG can’t go any higher. I am going to say though that it’s overpriced given the cards that are showing. Peter Lynch compared the stock market to a game of seven card stud, where every new piece of information released about a company is like a new card showing. The more you find out about a company, the more accurately you can gauge the value of a company, and thus whether it’s worth buying. The problem with Google (and similarly with Apple, but to a lesser extent) is that they try very, very hard not to release any information. What’s worse is that Google has built up a mystique about itself, so that any little piece of information (like an otherwise irrelevant partnership with Sun) gets turned into wild rumor and speculation (like that Google will make a web-based Office suite). If you believe the speculation will come true, that means that it will effectively break one of Microsoft’s monopolies, thereby entitling Google to its current market cap. But that means that Google has to deliver on something they’ve never even hinted that they’re working on. If Google’s going to keep its stock from going through a “correction”, it better start showing some more of its cards, and they better be good. Otherwise, they may end up like another company that valued secrecy and was going to take down a different monopoly.

MJ Ray in Will the last ex-UEA-er to leave Norfolk please switch out the light? asks why lots of people are now leaving Norwich... In my case, it was fairly simple, I wanted out of Paston, I hadn't really had time to look around as such, so when JD suggested that there was a job going at Runtime and would I be interested, I kinda jumped at the chance, threw across my CV (which still badly needs editing), and got an interview. As soon as the results where in, and I knew I had the job, I handed in the magical letter of resignation to Paston and moved my arse down to Brighton. It's a much better job, with more fun things to play with. Asterisk has been rocking my socks recently, and we should now have an almost fully functional set of phones on the asterisk pbx in London (which I'm maintaining from Brighton \o/). Anyways - it's nearly the weekend, and this time I *WILL* finish off reading the licences and throw Myon a mail (promise!).

I've gone from managing two buildd machines a few months ago, to five now, six or seven soon. Whew. New machines are bob and wendy (armeb), and ska/kiivi (ska was set up around the time kiivi broke down; kiivi's been fixed now, so that means one extra box too). Future ones may be jazz (m68k; Quadra950) and a third armeb box. Getting pretty crowdy—but nothing I can't handle. Yet. Though it is the highest number of buildd machines I've ever managed before. Since Christian Steigies is back from holiday, he had a look at garkin, his atari-with-CT60-board. It had lost its serial connection while he was away, so it didn't build anymore. But other than that the connection was gone, the box was still alive and kicking -- just needed to download a few megabytes of updates, which takes rather long through serial. In case you wonder: there's no ethernet connection, because the ethernet boards haven't been manufactured yet (they've been designed and ordered at the factory which will create them, but that hasn't happened yet). So with kiivi and garkin now both building again, the graph is going in the right direction again since a day or two. It's even more visible on the second graph; if you have a look at the new quarter graph, you'll see that the level of up-to-dateness is currently at its highest level in weeks; I count 28 days between now and the previous peak that is at approximately the same level as today. Which is good, of course. I'm quite happy about that quarter graph, BTW. The usefullness of the other two graphs (one with all information since the statistics were being gathered, one with the information of the last two weeks only) was seriously deteriorating. What this graph shows is way more useful... now if only there'd be a quarter version for the other graph as well...

Breaking radio silence to complain about something both trivial and technical, so it’s probably best if everyone skips this post. Unless you write a CD ripper or audio tagger. Then please read this post, because I’m going to explain why you’re wrong. (Someone emailed me about why Quod Libet acts this way, and I figured I’d make my answer public, with the hope it helps things.) History
Decades ago, before there was Ogg Vorbis, people used an obsolete standard called MP3, and tagged it with something called ID3 (in particular, ID3v2). One of the possible tags was TRCK, for the track number. It was stored as a string in “x/y” format, for track number x out of y total tracks. Perhaps this could be improved upon; maybe not (I’ll make a case for not later). Regardless, that’s how hundreds of thousands, if not millions, of files were tagged. Ogg Vorbis Comments
These days we have something better — Ogg Vorbis and its tagging format. Vorbis specified a tag called tracknumber, analogous to TRCK.

TRACKNUMBER: The track number of this piece if part of a specific larger collection or album

Now, this is a bit ambiguous. In particular, TRCK says how you should store the total number of tracks, and the Vorbis comment specification doesn’t. A strict reading indicates you can’t, but then the same standard is designed to be particularly loose about the data stored in the tags. So a lot of programs, including really popular ones, just kept their ID3 format and wrote tracknumber=x/y. Everyone was happy. Then someone — I wish I knew who, so I could write this to them personally — decided trying to parse the track number in this format was too hard, and so decided they would make a tracktotal tag that stored, as a single number, the number of tracks. Whoever Did This Is Wrong
Whoever decided to do this was pretty dumb. There’s a number of false assumptions here, roughly in decreasing order of annoyance:

• When you make a new standard, people won’t follow it (assuming otherwise is one of the stupidest design decisions possible). Lots of programs still wrote the x/y format. Quod Libet still does. I think it’s a better format. So now all that happens is everyone needs to implement both formats.
• When you make a new standard, people won’t follow it (as bad as it is, they made this mistake twice). Anything that does Vorbis tags probably does ID3 tags, and ID3 isn’t going away any time soon. So you still need the x/y parser no matter what.
• When you make a new standard, people get it wrong. So now we all need to support totaltracks, tracktotal, and x/y.
• Every other new tagging standard uses something much closer to x/y. Apple’s MP4 metadata, for example, uses two bytes in a single trck value. APEv2 uses Track=x/y. This is pointless differentiation.
• x/y is not hard to parse. There’s a horrible function called atoi that does it, and a lot of better ones.
• Using a simple numeric string still needs atoi and friends, because you don’t want “2″ to sort after “11″.
• You’re just trading parsing for reconstruction later. Quod Libet displays “Track [tracknumber]” which gets it right every time. Otherwise you have two strings “Track [tracknumber]” and “Track [tracknumber] of [tracktotal]” (no, you can’t just have “of [tracktotal]”, that ruins i18n).
• Users would rather enter a single “tracknumber=1/21″ than “tracknumber=1; tracktotal=21″.
• All else being equal, tracknumber=x/y is shorter. Maybe not a big deal, until we see a proliferation of discnumber (requiring disctotal, and totaldiscs), partnumber (requiring parttotal, and totalparts), and …

The result is that every program still needs to understand x/y — even if every Vorbis file used the new format — and the new format, and also needs to do the work to construct track numbers in the new format, assuming it wants to use them. That’s why Quod Libet converts everything to x/y format. I urge authors of tagging applications and CD rippers to abandon stupid standards. Instead, you should be spending time fixing your ID3v2 readers and writers, because they all suck too. P.S.
If anyone knows what program is putting ISO-8859-1 values into Vorbis tags, please let me know. I have spiked cluebat for you. A Summary of Other Things
I moved into a new apartment building next door to my old one, owned by the same student co-op, so it doesn’t actually seem like I moved at all. School started and is going well. I saw DOOM earlier this week for free, and enjoyed it. I’ve learned a lot about Python i18n over the past few months and am writing a brief howto explaining how to do it properly.

You should be good text file get at a place as we set window type a different countries, and fun way of code which is said I'm a drink! We point and displays it can be to the server and Free Software weblog at Debconf this has a place as a working at this unusually cheap used Google AOL, real: post Microsoft Apple hype going to be rid of the second one internal structure, explained how that he'd been a few more important. Ian Murdock Developers! As if I needed to take the man claiming to look at some relevance and he said the test cases.

Obviously this before and a bit of Launchpad team: have several staff who didn't plan to finally on the code, we want it.

Working with serial lines can sometimes give you big headaches. I have an embedded PC based on 386 SX 40 processor. This PC doesn’t make much but it has programs using the serial line intensively. Things didn’t work as well as expected so I looked carefully what was going on … the beast was loosing bytes ! My information has been promptly confirmed by the /proc/tty/driver/serial entry. If you have “oe: X” (where X is a positive number) there, it means that one of your UART detected overrun errors. So what’s an overrun ? An overrun happens when the UART receives data while its FIFO buffer is full. Why is the FIFO full ? Because Linux didn’t treat the serial interruption quickly enough. Why is Linux so slow ? Linux is not a real time OS and it doesn’t guarantee any response time to interruption, so Linux is not so slow but my PC really is … what happens is that interruption related to the network are treated before serial interruptions. Furthermore IDE disk interruptions can take too long too. Worst case is of course, you’re treating a disk interruption, then you have to treat the network interruption and only after that you can treat the serial interruption which in fact happened right after the beginning of the disk interruption So fixing serial overrun is a rather complex problem since it’s really a kernel related problem. Googling on the subject I have found several ideas to explore :

• configure IDE disk to use DMA hdparm -d 1 /dev/hda, use of DMA will shorten the time where IRQ will be masked to the kernel (in my case it doesn’t work since I’m using DiskOnModule which do not support DMA)
• make disk IRQ interruptible with hdparm -u 1 /dev/hda
• use irqtune to re-prioritize the IRQ on the interruption controller. This software is no more maintained and it doesn’t work out of the box on kernel 2.4.x.

Using hdparm -u wasn’t enough to solve my problem… so I continued to look for a solution and I found one ! I recompiled my kernel with the low latency patch and the preemptible kernel patch. Those are usually used for multimedia applications where you need good responsiveness in order to deliver content in real-time but the fact is that they do work for my purpose too ! My serial overruns are completely gone at 9600 bauds. However I can still have some when running at 115200 bauds. Moreover I can create serial overrun by running a find / -type f grep -v /proc/ xargs md5sum in the background… I can’t make miracles with this slow processor… if you have more ideas to further improve the situation, I’m willing to try !