DRI (3D acceleration) stopped working for me. The reason is simple (as revealed by looking at the Xserver log file): it was missing the DRI library for my graphics card.

So if 3D acceleration isn't working for you anymore, try this:

aptitude install libgl1-mesa-dri

This package includes the library required for DRI on my Intel 915 graphics board. Not sure how it got lost, though. Maybe I purged it when trying to remove "unneded" packages from my system (in order to free up some space)... I tend to try to uninstall everything I don't think I need.

P.S. Looks like you can now but ATI/AMD graphics again, given that they're releasing specifications for their chips - reliable opensource drivers on the horizon!

I've fought both the Nvidia and old ATI driver hell - not having to do that was one of the main reasons why I wanted Intel graphics. Now ATI/AMD is back on my radar.

For my diploma thesis (on clustering of high dimensional data, especially correlation clustering, i.e. clustering data by properties such as data correlation) I needed an easy way to filter out data from CSV files as used by gnuplot (actually, they are whitespace-separated).

I hacked together a small tool that would allow me to easily 'grep' out certain parts of the datasets. If you know of a similar tool, please send me an email to erich AT debian DOT org.

Currently, the tool allows you to do commands such as:

gpgrep "1~5$" "1<200" < out/3d-2lin-noise.variances

The tool is still in early development. Syntax may change, and I guess I'll add some more filters. But you get the idea. Focus is on a very compact syntax.

A future filter I'm considering would e.g. select 20 random rows (with a fixed seed value, so it's reproduceable!) for sampling and a modulo-match to select e.g. every 7th row.

Maybe I'll also add some output processing later, such as averaging values, calculating variances and mean deviations, stripping away columns (but you can do that in Gnuplot already). I don't want to overdo it though - it's just meant as a mini filter you can add to script output visualization. It's not meant to replace a full statistics toolkit.

So if you know of a simple tool that can do that already, please tell me.

P.S. a few people have pointed out awk. Yes, it can do most of this. One thing I also need, and I'm not sure on how to do that the easiest way in awk is to preserve 'blocking'. That is empty lines. Because the data set

1 1
2 2
1 2
2 1

Ich arbeite zur Zeit (neben meiner Diplomarbeit) auch etwas für Amiando, einem sehr praktischen Einladungs- und Ticketservice. Die Grundfunktionalität des Dienstes ist kostenlos, und sehr praktisch z.B. für Geburtstagsfeiern, oder wenn ihr bei irgend einer Party eine Gästeliste benötigt. Kosten entstehen erst, wenn ihr Tickets verkaufen wollt.

Verzeiht mir also mal diese 'schamlose' Eigenwerbung hier. :-)

Als Werbeaktion bietet Amiando gerade ein interessantes Gewinnspiel an, zu gewinnen gibt es nämlich:

• 2 Nächte in einem Münchner 4-Sterne-Hotel für 4 Personen vom 5.-7. Oktober
• + 500 € Taschengeld zur freien Verfügung
• + 1 komplettes Trachtenoutfit (weiblich/männlich)
• + 4 Karten für das ausverkaufte Derby Bayern München gegen den 1. FC Nürnberg
• + U-Bahn Karten für 4 Personen an allen 3 Tagen
• + 4 Gästelistenplätze im exklusiven Promi-Club P1 in München

Dafür müsst ihr nur eine entsprechende Einladung erstellen und vier Freunde einladen - und möglichst kreativ in der Einladung beschreiben, wie ihr die 500 € auf den Kopf hauen werdet! Sprich: einfach mal den (kostenlosen - keine Bankverbindung oder so was notwendig bei der Anmeldung!) Einladungsdienst ausprobieren. Die Benutzeroberfläche ist nämlich ziemlich überzeugend. :-)

Derzeit sind etwa 17 Events in der "Oktoberfest-Gewinnspiel"-Gruppe eingetragen, und mindestens einer heisst noch "Eventtitel hier eingeben", - die Chancen sind also ziemlich gut. :-)

Details zum Gewinnspiel

Debtags, a tag-based approach to classifying Debian software packages, has taken another big step forwards.

Debtags was included in the relaunch of packages.debian.org, Debians package search and information web server. This means it's now visible to pretty much any Debian user.

The experimental packages.debian.net also uses Debtags to recommend 'similar' packages.

There are some AI/Datamining projects around Debtags that I'm interested in, but I don't know when I'll find time to work on them.

I spend a few hours today trying to get Scalix to work right...

I managed to be able to loginto the admin console at the end, yay!

Still: it seems that it's pretty much unusable except by Outlook (maybe, I havn't tried that).

The reason is simple: they don't let me log into the web mail. With none of my browsers. Because they havn't understood what 'Gecko' is.

For tech people (apparently the Scalix people don't fall into this category), Gecko is the 'rendering engine' inside Firefox, Mozilla and a few others.

With 'others' including the trademark-unencumbered Iceweasel fork of Firefox, or the Gnome webbrowsers "Epiphany" and "Galeon". Which are not "supported" by Scalix (do you happen to actually know what you are supporting? Did you develop for the Gecko rendering engine or for the Firefox UI?).

Other people in the "I'm a web 2.0 company, but I don't know my stuff" hall-of-shame include the huge german email provider GMX. Their new webmail 'preview' is also only enabled if you claim to be using Firefox or Internet Exploder. They neither have understood why all Mozilla browsers include the Gecko id...

Many people will have seen their postfix reporting new errors these days, like if is overloaded:

postfix/master[7287]: warning: service "smtp" (25) has reached its process limit "100": new clients may experience noticeable delays

This is not really because your system is overloaded. To me, it looks more like a (deliberate?) malfunction of some spammers software. When you rejected it's spam mail a few times, it starts opening up lots of connections to your system, but never sends any data. They just sit there and time out.

The solution is simple: limit the number of connections per client. Make sure the 'anvil' service is enabled (default at least on Debian) and then add

smtpd_client_connection_count_limit = 10

You should now begin to see messages such as

postfix/smtpd[21680]: warning: Connection concurrency limit exceeded: 11 from unknown[210.210.247.52] for service smtp

A corean cable IP. Yeah, I don't think this would have been legitimate mail traffic anyway.

Feel free to pick the limit value even lower than 10 on really busy systems (or if these 'attacks' become stronger) - most people should never need more than 2 parallel connections (note: you can send multiple emails over one connection, there isn't much to be gained by using more than one connection).