puzzling.org

I hear about software rotting a lot, but not a lot about hardware rot. But rot it does. No sooner had Andrew and I got back to Australia about this time last year (354 days ago, actually) and slowly but surely coaxed our desktops back to life over the course of a week or so of patiently reseating cables and whacking them hard on the side, it seemed, did we moved house (five and a half months ago) and somehow destroy them again.

Being laptop enabled, I didn’t find this out until I wanted a Windows machine to do my tax return. Booting my old faithful produced nothing more than a loud and upsetting pop. My intimidating hardware acumen lead me inevitably to diagnose some kind of bang in the motherboard thingie. Or actually, probably in the power supply, because it smelled weirder. (On a tangent, I still count the day that I showed my father how to plug the CD-ROM back into the motherboard and the adventure went to his head and he ended up taking the case’s power supply apart and carefully vacuuming it as one of the odder experiences in my life. Thanks for the memories Dad.)

So then I tried Andrew’s slightly less faithful desktop only to find that it booted into a completely broken Windows 98 install by default and couldn’t be fixed because it didn’t work with PS/2 keyboards anymore, so the BIOS is uneditable.

So I did my tax return on paper and resigned myself to using up perfectly good airspace in our house by filling it with completely useless computers (I’ve never worked out how to dispose of computer bits). And that was all fine and good until Andrew’s laptop was stolen and we were down a computer, a key element of our lifestyle. Surely between two broken desktops we could put together one working machine? Unfortunately, neither has such an accessible setup anymore: my BIOS is kaput, his is not editable. But we eventually did it by shoving the hard drive in another machine.

Incidentally, a big boo-hiss to Ubuntu for requiring a key-press to boot their live CD (apparently, Andrew was the one who waited for it, but not me). Andrew’s BIOS doesn’t like USB keyboards, so he needed an operating system that could be booted with no key-presses.

As most of the relevant community already knows, there’s a O’Reilly Twisted book out. You can also buy it from an Amazon associate to help defray the twistedmatrix.com hosting costs.

An important question is being discussed as I write: what should the book’s code name be? If Programming Perl is the Camel book it is argued, this should also be coded on its cover. So far, the orgy book seems simplest, with the snake orgy book as a backup if O’Reilly gets a bit wilder with their covers.

I’ve had several discussions recently with people interested in the why-so-few-women-FOSS-developers problem (pimp: it’s a problem talked about a lot now, got links for the bibliography I maintain?) specifically as it relates to LinuxChix. The general starting position is this: where is LinuxChix in the creation of Free Software? Well, while individual members are here and there, that tends not to have much to do with their participation in LinuxChix. LinuxChix is a user and social community, and further, it doesn’t seem to ‘graduate’ a lot of people into the big bad world of FOSS development. (Note that some chapters, particularly LinuxChix Brazil, operate pretty differently.)

This has come up in a few places online. I don’t know if Fernanda Weiden was thinking about LinuxChix when she wrote this, but it’s a good match for some of the more negative opinions:

That’s the role of the women’s groups, to offer a friendly interface for women to get their feet wet and then join the community. The problem is when these groups don’t have a clear target, in the end they turn in Barbie worlds that don’t exist in reality. Instead of integrating the women into the community, they serve as ghettos, re-creating existing groups in the community with the only objective being more friendly for women.

Máirín Duffy writes:

LinuxChix gets a lot of mention in the essay is referred to as being an open source development community, but I feel quite strongly that it is not. Some of the motivation behind my pushing for GNOME Women was borne out of frustration with LinuxChix. LinuxChix is really more of a Linux User’s Group (LUG) than an actual development community.

I raised this on a LinuxChix list today and got an interesting response from Carla Schroder. To paraphrase greatly, the upshot was that Carla draws a distinction between not being an open source development community, which pretty much everyone agrees that LinuxChix is not, and not being an activist community.

LinuxChix is pretty broad brush, but some of the things it is active in are: providing a forum for answering technical questions from women and providing skills education to women. And that’s only the more formal stuff. Behind the scenes, there’s a long tradition of discussing technical careers and related things (interviews, salaries) and working on giving women recognition for their technical accomplishments (within the community, mostly). Carla also pointed out that while there aren’t masses of FOSS developers emerging from the community there are quite a lot of women technical writers and a huge number of women sysadmins who’ve derived a substantial chunk of their career launch from LinuxChix help. It’s not a ‘ghetto’ in that sense.

LinuxChix is not active politically, and it’s not a development community. (It helps coders, but it doesn’t have a coherent project.) It does suffer from being seen as the one-stop-shop for women-in-FOSS when it’s actually not doing some things that women would like to do and it’s particularly not doing work to produce women FOSS developers. But that doesn’t mean it should be mistaken for a sheltered workshop, it just means there’s a difference between reality and publicity that could use some work on one direction or the other.

Andrew and I have been talking about what makes a good technical community Planet. Here are some thoughts about what the best planets have, keeping in mind that it is probably not possible to achieve this by any other means than luck:

Critical mass

This seems to be somewhere around about the twenty or thirty writers mark, just in order to achieve content that changes regularly. This seems to be the major failing of Planet Twisted. You might get regular content from as few as three or four people, but it will get same-y.

Community leaders

It’s most interesting to see actual decision making discussion happening through entries themselves, but after the fact analysis is nearly as good. But you won’t see any of this if you don’t have key decision makers or at least opinion makers blogging about decisions and direction. I see this as one of the big advantages Planet Debian and Planet GNOME have over Planet Ubuntu. My primary interest here is for interesting reading, but this has a bonus side-effect of greater community transparency: a lot more relative outsiders scan Planets now than scan development mailing lists.

Expert writers

It’s useful to have a significant set of regular writers who write in some area of their personal expertise, be it user interface design, translation, programming or cat herding. This decreases the sense that you’re reading uninformed and predictable reactions to technical decisions and increases the chance that you’ll learn something new by reading.

It adds a bit of spice to have some of these writing about project irrelevant expertise (climate change, workout programs, photography techniques) because it increases the “smart people who I’d love to go to dinner with” vibe. It can be overdone though: while it’s likely that your writers are the best there is to offer on technical subject X, it’s less likely that they make as universally interesting reading on their other interests, because it’s less likely that they’re the people worth listening to and it’s less likely that their interests overlap with those of the readers. Sometimes you get lucky and sometimes you don’t.

‘Journal’ blogs are hard to pull off on masse: it’s both a tempting format and a difficult one to make work for readers who have no prior interest in the writer’s personal life (ie, everyone who isn’t friends or family). These blogs are much like outside interest ones: considerable payoff in community spirit if they work, risking uninteresting reading in many cases.

A Planet community

The best Planets are full of people who actually read the Planet and let it inspire their entries. I also find that planets from a focused technical community have an advantage over planets from a user or random community because they’re more likely to make for coherent reading with common threads.

Inter-entry conversations

Once you have a Planet community, the best Planets start having extended inter-blog discussions during which you can see opinions being formed, revised and finalised. This is the Planet zenith.

So-called memes (standardised entries based on quiz results or a common entry template, which tend to be quick and fun to write and therefore highly contagious), as in all other blogs, generally are the most common and most boring kind of inter-entry conversation. There’s not a lot to gain as a reader from knowing that someone is more like Bilbo than they are like Sauron or which countries they’ve visited unless this is used as a platform to make some kind of debatable point (almost every opinion that’s actually interesting is debatable) or to tell a story, and people do this too seldom.

I need to pay more attention to DebSIG I think. How did I miss Conrad Parker on What can Free Software teach Free Media? This is what comes of being the only geek in Sydney who doesn’t like beer or pubs, I guess.

del.icio.us is rather old hat really, but I wanted to copy my comments here from dria’s blog about why I like it.

For those who haven’t heard of it, del.icio.us is a bookmarks site. Once you create an account, you can add a little pop-up button to your browser toolbar, and when you see a site you want to store for later reference, or recommend to someone else, you click on it, it loads up a simple page for you to enter descriptions and tags, and then its stored. You can get to it from anywhere.

It also has a so-called social software component in that it shows you who else has bookmarked that URL thus potentially introducing you to new people who have the same taste in links as you do. See for example [people linking to me]. (Why md5 for the URL hash, I wonder?)

Here’s what I like about it:

• a fairly flat site for bookmark storage where things are tagged rather than buried in folders. I can find my bookmarks more easily on del.icio.us than I can in a web browser’s bookmarks menu;
• bookmarks accessible via my del.icio.us user page from any web browser (the Hotmail advantage); and
• the handy provision of the ‘remember this’ bookmarklet so that I can click “pop up post to del.icio.us” in my taskbar at any time and post a link to the page I’m looking at for later reference.

I’m actually more or less neutral about the social aspect of it. Links tend to divide into three categories: links chosen by several hundred users; links chosen by a few users and links chosen by me. There doesn’t seem to be a query engine that allows anything more sophisticated than who else linked to this one thing? I can see myself asking who links to things I’ve tagged politics? or who links to some of my less popular choices? as ways to find bearers of fresh links, but the fact that we have one chosen link in common is uninteresting to me.

Related links:

• de.lirio.us, like del.icio.us but Free (via Beth Skwarecki); and
• CiteULike, like del.icio.us but for academic citations (via Liam Hogan).

I’m really enjoying Bazaar 2.0 as compared to GNU Arch. There’s a lot of things I could say about Bazaar 2.0 as version control, but let’s leave it as ah, that’s why I originally liked the idea of distributed version control.

Now that that’s out of the way, I have something more pressing to communicate! Bazaar 1.0 has a commandline program baz, pronounced like the first syllable of bazaar (I note in passing that until I saw that commandline, I thought that it was called ‘bizarre’ and am still disappointed). Bazaar 2.0 will have bzr. Now, people, you can’t seriously be expecting me to pronounce the latter as bazaar can you? C’mon, … try saying it. bzr. bzr. BZR. It doesn’t quite have the Australian neutral vowel in it, but I’m determined and ready to fight for it to be pronounced buzzer. You can have all kinds of really nerdy jokes about hitting the buzzer. Now get with the program.

My theory on this is that you should:

• pick an implementation language;
• pick all the surrounding software (mailing list manager, revision control system, CMS for the website if necessary);
• write and release something that works (or is pretty).

Only then do you indulge in even the smallest bit of community building.

Then you can avoid the six month long startup argument about which tools to use to write the vapourware. These arguments even trump the ‘what features should we do first’ arguments!

Rusty Russell is not happy about John Quiggin’s embrace of Creative Commons’ Attribution-NonCommercial-ShareAlike 2.1 Australia License as a kind of a good default for allowing other people use of your creative work.

Russell slams Quiggin:

That Quiggin takes this path despite training as an ecomonist [sic, original author’s emphasis] demonstrates either a lack of deep thought on this issue, or that he uses economics to justify his leftist dogma, rather than to examine issues. (This paragraph was about as polite as I could make it).

I can’t say that I’m too much of a fan of the phrase “leftist dogma.” It’s about as meaningful to me as saying “fropbutz dogma” — ie I tend to prefer attacks on political positions on a particular issue rather than attacks on them because of other political positions that they’ve been known to be associated with. (What are the pragmatics of the word ‘leftist’? I read it as having exclusively hostile connotations, that is, that a position or group of positions is only described as ‘leftist’ — as opposed to ‘left’, ‘socialist’, ‘communist’ — by people who oppose it.) To be fair, this is what the rest of Russell’s piece does, I’m just having a go at the ‘leftist’ ending.

But that wasn’t what struck me enough to write an entry about it. I was struck more by the emphasised part: “despite training as an economist”. This strikes me as a cutting line indeed. Quiggin’s relationship to economics isn’t that he trained in it: it’s that he is an economist. Russell is implying that his economic positions would be foolish from someone with an undergraduate major. It would be like saying that Russell is a remarkably bad coder for someone who’s met a few kernel developers.

Thanks to Linux Australia we have some grant money for the Sydney Twisted sprint being held in two and a half weeks. Check it out if you’re coming:

• airfare subsidies;
• accommodation subsidies; and
• a sprint dinner.