Raw Matter

I am subscribed to the Google news alerts for the keyword Debian, and usually this means that I am notified of every security upgrade at least three or four times (needlessly, since I'm subscribed to Debian's security announcements anyway...) But today, two headlines sparked my interest:

One, Carla Schroeder wrote an introductory article into building Custom Linux kernels with Debian and Fedora (though why somebody would want to do this is beyond me...), showing how easy the process is (I may be biased, but I find Debian's way is easier. On the other hand, it within the -- rpm -- package framework instead of a custom command -- make-kpkg in Debian -- feels more consistent and logical. Without reading documentation you won't get anywhere on both systems, so it probably doesn't matter...) One down for Carla, however, for not doing enough basic research: “The new naming convention [linux-* instead of kernel-* packages] opens the door to using other kernels with Debian, which I don't know what those would be, but Debian is ready for them.” Or one down for us, the Debian project, for not making enough noise around the various non-Linux ports. Though that will come when GNU/kFreeBSD is released with lenny...

Two, Debian derivative SimplyMEPIS, which switched to Ubuntu as a base operating system a year ago, is now returning to Debian GNU/Linux as its base system. The main reason seems to be long term stability and upgradeability - which is a key strength of Debian for many other users as well (me, for one.) The reason for switching to Ubuntu in the first place has been newer application versions and predictable upgrade schedules (and that Warren Woodford misunderstood how Ubuntu LTS was to be maintained, apparently) -- both issues are being addressed within Debian as we get better release management and faster integration of new application versions (the big pieces of software I use regularly are the Linux kernel, X.org, KDE and OpenOffice.org, all of which I find are integrated in unstable fast enough, and much faster than 2 years ago. That the testing migration sometimes takes time is not a problem since mixing etch, lenny and sid is usually possible without problems even in a production (desktop) environment.) So, if we take a step back from the day-to-day squabbling between developers working at cross purposes and look at the output we produce, I believe we're still doing some things right.

Russel Coker cites evidence that the availability of Pornography may decrease rape incidence. I admit that I haven't really read that article myself, maybe the topic is addressed: I frequently hear that parts of the porn industry are mixed up in drug trafficking, slave trade, money laundering etc. So I'm not quite sure what these results signify, or what action could or should be (or should not be) taken as a result.

The recent announcement that sparc32 is to be retired seems to be regretted by many. Various suggestions on what should or could be done have been put forth. Here's mine: let's at least create some minimal project infrastructure for a (hypothetical) sparc32 port.

So, there now exists a sparc32 Wiki page and a sparc32 mailing list. All that remains is people who actually do stuff.

I recently learned that HP updated their Debian compatibility matrix to include Debian 4.0 a.k.a etch (which is obviously great, and I hope more major vendors would include this type of compatibility information.) Clicking around, I find it somewhat funny that if you go to the certification information for Red Hat or Novell SuSE (now there's a great name slowly dying...), you'll find a Debian logo, a Red Flag logo and a Mandriva logo. But no Red Hat or Novell/SuSE logos. Do they want their customers to try Red Flag, Mandriva or Debian instead of the Big Two(tm)? Because on the Debian page, there's no Red Hat or Novell/SuSE logo...

Just playing around with powertop and laptop-mode. Why didn't anybody tell me about the excellent documentation of /proc before?