Resumé

This year's meeting was the best I've ever had. The reason could be that this is the first meeting I was able to attend properly, but I rather believe that it's because of the very good mixture of different people, architectures and work as well as the proper size (about 20 people) of attendees. I have had very much fun (less sleep though).

We have used nearly the entire room. I guess that at most ten more people could attend without having space problems. Even though we have had quite a lot of machines (ip numbers range from .1 to .74) we didn't have a single power outage. This I don't understand, did the university change their fuse since the last meeting? We've transferred something like 2.5 TB of data during all four days. It was limited to the 10MBit/s connection to the University LAN mostly, let's find out what will happen when we use 100MBit/s next year.

Having external dinner on all three evenings also revealed lots of fun, resting from real mens work for a while. However, having dinner with approx. 20 people doesn't seem to be easy for some restaurants, we will have to train them more, it seems.

Many thanks go to the mathematics department and the computing center of the University of Oldenburg, who made the meeting possible.

Architecture: mips/mipsel

The sysmips problem for R3000 processors seems to be fixed finally. This was keeping all R3000 based machines from being useful since random programs segfaulted. Newer R4000 based machines didn't have this very problem, but not everybody has gotton such a box.

The addinitrd tool which is used to append an initial ramdisk to a formerly compiled Linux kernel was fixed to work again. Without this one had to manually convert the ramdisk to an object file and link the kernel again. This tool was tested on MIPS but should work on MIPSel as well.

Boot-floppies and working Kernel packages (out of which boot-floppies were built) were uploaded after a bug in kernel-package has been fixed. This was the only thing that was missing from mipsel being considered as architecture for Debian woody.

Architecture: hppa

Graphics on an HP 9000 712/100 now works with correct colours. X ran on a framebuffer already, but the framebuffer wasn't able to set the colors correctly, so it was interesting to work on this machine.

Next meeting?

The next meeting shall take place at the last weekend of September 2002. I'll negotiate with the University of Oldenburg again to be able to use the same rooms like this year, since they were properly suited.

... to be continued...