April 23rd, 2005
| Filed under Misc
|
Eben Moglen talked this morning on the current state of freedom. Much interesting stuff, and it looks like he’s making progress. In his opinion the SCO debacle is over now, and some gains have been made in other areas, including overturning the FAT patent. He got two standing ovations.
Attended Anton Blanchard’s talk on bringing up the Power5 architecture. They can do a 1-second kernel compile. Impressive. It remains to be seen if the Power5 makes it out of main-frames though.
We’ll probably attend one of the afternoon sessions and head off then to get home at a reasonable time.
April 22nd, 2005
| Filed under Misc
|
More heavy stuff today. The keynote was by Andrew Morton. This was not what I expected; Andrew does not so much do magic as just work very, very, very, very hard. A constant stream of patches come past him, and he addresses every single one, even the stupid ones. This is just mind-boggling.
Tridge did a surprise extra talk on the new LDAP-like database system for Samba4. Nothing stunning here, but Tridge’s talks are always full of useful little snippets and facts you end up using somewhere.
The main thing in the afternoon was Keith Packards talk about re-architecting the X Window system. This was not just about technical changes but political ones two. But he also did the demo of the wobbly-windows on the GL based server.
Also went to the related Cairo talk. Much useful stuff, and Nick managed to give Carl Worth and Keith Packard pointers to useful information on tessellation algorithms.
Also did the kobjects talk (heavy but useful to know how sysfs works internally) and Knoppix (actually walked out of that one; too high-level and anecdotal for me). It’s the ‘Penguin Dinner’ tonight; should be interesting.
April 21st, 2005
| Filed under Misc
|
I’m going to try doing a running update today …
Tridge did the keynote today; he was mostly talking about software engineering tools in the opensource world, and the advantages they give. His main point is that the sum of all these tools is greater than the parts. There’s a few things there I can probably apply to refactoring the AG AV tools, especially valgrind, gcov and talloc. The other interesting thing that I wasn’t expecting is that he addressed (briefly) the BitKeeper issue by demonstrating how simple it is to access BitKeeper repositrories. You’ll probably see the following code about a lot over the next couple of weeks:
echo clone | nc thunk.org 5000 > e3fsprogs.dat
Attended the session on Netem, a system to simulate large network systems (specifically latency and packet-loss).
Went to a session on Linux Kernel Scalability. In practice it was an overview of the problems and solutions to locking in the kernel. Seriously in-depth stuff, and it lost me at a few places, but it did include some discussion of the RCU implementation in the kernel.
Attended the Asterix session. It was more a tour of the advantages and what you can do with VOIP. Not bad, but I was hoping for something more in the cool-hacks vein.
Attended the delegates mixer/networking session. This was held in the CSIRO discovery center, which is a kid’s science center so was kind of fun after a few beers.
April 20th, 2005
| Filed under Misc
|
Tuesday was mostly spent in various GNOME sessions. One interesting thing is that the word collaborative came up. In particular Jeff Waugh mentioned that some of the Abiword hackers are working on a collaborative version of Abiword.
Today was entirely spent in the Linux Kernel hacking session. There were a few other things I wouldn’t have minded seeing but I figured just about anything I learned in the kernel session would be useful sooner or later, and Robert Love and Rusty Russell are too entertaining to pass up. It was very intense though, almost entirely practical (we developed a driver for a virtual crypto device on Qemu), but I did get to debug my first kernel oops (having caused it in the first place).
April 19th, 2005
| Filed under Misc
|
Myself and Nick are spending the week at linux.conf.au in Canberra. Although the conference proper doesn’t start until Wednesday there are mini-confs on Monday and Tuesday so we headed down yesterday morning. Most of they day was travelling and settling in (aka ‘debugging the wireless connections’), but we did manage to catch part of the Debian miniconf and the panel on linux clustering (to our suprise included Youhzen from AC3). We should manage to get some more in today though, including the EduLinux talk by my old dotcom stable-mate David Peterson, who is going to get heckled to hell and back if I have anything to do with it.
