Older blog entries for trs80 (starting at number 37)

This is actually a reply to a post by Jono Bacon of LugRadio fame, but his comment system doesn't like me, and it's worth repeating to a wider audience.

Free beer and speech: You should never, ever, use the terms free as in beer and free as in speech, as they're opaque to anyone who hasn't read the specific piece[2] by RMS that defines them. For a long time I had them the wrong way around, since it made more sense to me that way - I'd heard of "open source beer"[1], its recipe having a free license (RMS himself uses recipes as examples of Free software), whereas the GFDL (for speech) allows invariant sections[2], and a lot of RMS' writings are freely distributable, but only if unmodified.

[1] Although I'm sure I heard of it before 2005, when Vores ├śl was announced.

[2] The first distro I used was Debian, so I'm more influenced by the DFSG than RMS' Free Software Definition.

the inside of an aircon with green corroded copper pipesInstalled a new aircon at UCC yesterday, to replace the dead gobi, seen at right, after we pulled it out. Mmm, corrode-y. We named the new aircon gibson (all aircons and friges are named after deserts) because it has an IR remote, and future interfacing will result in ... well, you can guess.

Spent Thursday, Saturday night and this afternoon upgrading manbo which is now running Solaris Nevada b28 and is fully operational with 14 400MHz UltraSparc II CPUs, 12.25GB of RAM and 216GB of ZFS, so UCC should be opening it up to FOSS developers to dtrace their apps and tinderbox on. This required quite a few reboots, each of which printed "svc.startd: The system is down." which triggered the Strong Bad Techno in my head each time.

netmon-applet ate my panel, yay for %gconf-tree.xml editting. f-spot seems kind of cool, when it's not crashing. Will hopefully get to work on weather.gnome.org while on holiday.

Work: Set up a 2.55TB RAID array using the Areca ARC-1130 controller and 8x Western Digital Caviar RE2 400GB drives. Had a few teething problems, including Ubuntu not turning on CONFIG_EFI_PARTITION and some odd interaction between the controller firmware and the drives where it only works from a cold boot, it hangs in the BIOS after a warm boot. However Areca support is great, they got back to me at 6:30pm on a Friday afternoon. This guide to LVM recovery was also useful for getting back my data from the clutches of truncated partitions.

UCC: One of the aircons died, and it seems to have caused two drives in manbo to die from overheating. Ebay has 5kW split systems for $550ish, maybe we should grab one of those. Also need to start organising the camp, committee has been very slack this year.

Advogato: Thanks to raph for restoring my account. The only niggle at the moment is the diary ratings - my explicit ratings are showing up, but not any calculated ones.

ncm, bagder: Why don't Windows people help with Free Software projects? I suspect the reason is more practical than what you suggest: developing software under Windows is PITA. A few weeks ago, I spent a couple of hours trying to set up a Mozilla dev environment on Windows (trying to help fix 244770, and failed miserably. I already had Visual Studio and Cygwin installed, but needed the latest library updates, the right directories in %INCLUDE% and %LIBS% and in the right order etc. When I compiled Mozilla on Linux (to fix 128398) I just untarred, configured and built. Admittedly Mozilla is not your average FOSS project, but the point still stands:
Linux: comes with a build environment ready and working.
Windows: requires paying for Visual Studio (or juarezing it) and often installing FOSS tools for development and general futzing around for non-native (ported from posix) apps.
Mac OS X: (badger's comment about "windows dev/user ratio lower than on any other OS") comes with build environment and dev tools (Xcode and the usual FOSS suspects).

Stuff: Need to: look at Bacula for UCC now we have tape library; organise intervarsity LAN with MINCS where much ET shall be played; finish coding Excel/VBA monstrosity for SQM assignment by internal group deadline of tomorrow evening; stop playing around with Asterisk at work and actually put the new PBX into production; work on weather.gnome.org; get daniels to fix 10866 so I can use my new laptop in X; sleep.

It seems Google has no sense of humour these days (I assume it had one in the past). First it puts the kybosh on AdBar, then it removes Greg Duffy from its results for writing an article on how to extract entire books from Google Print. Yes, both these violate their ToS ... but with the kerfuffle about AutoLink in the Google toolbar and who controls a webpage, what's to stop a user doing what they want with data that Google freely serves to them? (And yes, I know that argument is specious, but it's a good starting point).

In other news, good to see that there's a lot of discussion going on about MoFo and Seamonkey 1.8 and Firefox reviewers. I would like to claim to have been the person to start the 1.8 flam^Wdiscussion ;-)

Skipped a lecture to fix something at UCC this week (morwong just wanted a cuppa tea, a Bex and a good lie down) for the first time this semester, but it's only week 2. Which isn't that promising - I need to stop spending quite so much time in the clubroom, particularly at night. Also need to find a supervisor, 4th year project and write a proposal for it in a week, after the last two lecturers I sought out couldn't take me on. Gnome 2.10 release party tonight, will be nice to relax in between the many fresher welcomes this week.

Right, apache2 has some awfully useless error messages. Well, error message singular, because for a wide variety of error conditions error_log reads "Premature end of script headers". Which is pretty useless when the real problem is that you're using suexec and the file is g+w, or is a symlink. Although I must admit I only just now thought of checking suexec_log, which does have sensible error messages.

Davyd played me some LugRadio which is truly awesome, and fills a much needed hole left by Geeks in Space, which dates from back when /. was still even mildly cool. Contains some great lines, but none that sound good quoted out of context.

Anyone got any economic/finance bloggers to recommend? I was reading the business section of the paper today, and realised there was a lot of commentary in there, and that there would probably be bloggers covering the same issues.

Finally got the $10 power dongle to go in UCC's rather more expensive and chunky new dual opteron server martello. Found out Ubuntu doesn't have a AMD64 PXE installer, so had to use the Debian pure64 port instead. This went flawlessly, to the point where we could set up LVM on two degraded RAID 1 arrays (we haven't got the 8 port SATA controller yet) with no effort at all. However, once it came time to make the system bootable, m and I discovered neither lilo nor grub could support booting off an LVM volume on AMD64 - I thought we were past the days of requiring a separate /boot partition. So we gave up and went to get dinner and play poker and/or drink with many other Perth geeks.

yakk is back in town for summer, and is cool.

The above was written yesterday (the 10th) but Advogato didn't want to accept the post last night (and recentlog seems to be unusually short). Anyway, today was the usual Saturday UCC Basketball game (8 of us play in a social league), then back to UCC for the cleanup, which was very successful, and we even cleaned up the machine room for once. There's a quick before and after, or you can check out the full webcam archive.

I ran across an interesting post on Planet Mozilla talking about improving search by determining which emails were "interesting" to you. This could be done by monitoring how you read your mail - if you click on a link, it was probably interesting, if you only looked at it for three seconds, it probably wasn't. This is an good way to get metadata without requiring you to do anything, and could probably be expanded to documents you're working on, webpages you visited etc., maybe plugged into Beagle to choose the most interesting results for what I'm currently working on.

The hard part about a better search system is that generally if I'm searching for something, it's not going to be obvious, otherwise I wouldn't need to search for it. So if I'm working on a large manuscript (which would obviously rate as interesting to this system), I know where that is. But where did I save my shopping list that I wrote up while I was having a break? Or, (and this is my use case) what was that random (but now related) webpage I came across a week or two ago?

Slightly related wrt metadata, more metadata that feed aggregators should support.

Someone should set the goons on these safety images, hilarity would result.

28 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!