etfb

Someone just created an LJ community for weight-loss surgery people (users? victims? afficionados?) within the Society for Creative Anachronism. While I understand that the SCA is known for its blessedly relaxed attitude to modern standards of personal appearance (they say you can't do a nude calendar of SCAdians because they don't make paper that wide^[1]) it still seems an odd combination. Imagine the Venn diagram: there's a very small circle for people who plan to have or have had weight-loss surgery, and another very small circle for people in the SCA, and no doubt the intersection of those two circles is non-empty, but still: it's a funny combination.

It got me thinking. What if there was one forum, for everyone? You could self-identify by tags -- I'd pick SCA member, bard, SF fan, chorister, pedant, parent, home-schooler, cybernetic were-turnip, former president of Zimbabwe -- and then the only posts you'd see would be from people who matched some number of your tags. You could set it to "all", and get a very small but precisely-tailored community, or to different values of "some" to get progressively larger groups. It's the ultimate "us vs them" environment, except that it makes it clear, surreptitiously, that every one of "them" is an "us" in one way or another.

To make it interesting, you could have a huge collection of themes and styles, and let people vote on the appropriate style according to their tags: I could vote with my SCA member and chorister tags to have 6 point Comic Sans in shades of magenta and puce, and in any group that had those tags, that design would reign supreme... at least until more people with those tags voted for something else. The changes would be subtle for the "some" groups and more marked for the "all" groups, so you'd get visible feedback to tell you where you were. If you could merge themes in some way (colour scheme from here, font from there, images from somewhere else) then you could evolve the designs democratically, which would be interesting.

Could be interesting. One blog to rule them all, into which everyone posts. Like a merger of the friends page and the tag cloud. I wonder if anyone's thought of that before...?

(Edited to add: the domain taglog.com is owned but unused, but vennlog.com is available. If only I had time to work on this...)


^[1] Not sure who "they" are, but they're rude bastards.

The new ACTION bus timetables have been in effect for nearly two weeks now; poor miss_terri_girl has been hard at work telling the punters where to go, and the rest of us have been hard at work wondering why they shifted everything around without actually improving anything much. One of the things I've been missing anew, as a result, is the trip-planning tool that the stupidly named www.131500.com.au website makes available for planning your dalliances with Sydney public transport.

So I decided to make one.

The first trick was getting the timetable information in a form I can use. I saved copies of the timetables from the website, and wrote a Perl program to strip out the rubbish and reduce them to S-expressions, the format that Lisp understands. So a timetable like this two-tone monstrosity gets turned into an expression like this:

((15 315) 
 (:NORTH
   ("Tuggeranong Interchange" "Woden Interchange" "City Interchange"
    "Radford College Bruce" "Belconnen Interchange" "Copland College" "Melba"
    "Spence" "Spence Terminus")
   ((15 -1 -1 -1 -1 723 731 735 743 748)
    (15 -1 -1 -1 -1 803 811 815 823 828)
    (315 730 749 807 821 826 834 838 846 851)
    (315 830 849 907 921 926 934 937 943 948)

... and so on. This was mysteriously tricky, because it appears that the timetables are actually hand-generated by someone in the bowels of ACTION. No templated database generator programs here, no sir! The mind boggles.

Next trick is to work out how to use this information. Here are the questions I can imagine asking:

Where is my bus at this moment?

(reaction:bus 15 0803)

Given a bus route and a time, tell me what stops any such bus is currently at. For example, the above query should tell me there's a 15 bus at the Belconnen Interchange heading north, another between Woden and Civic, arriving at Civic in four minutes, and one heading south and just about to arrive at Melba, among others. This tells me whether I've missed my bus and what ones I can expect to catch next.

What buses are going past here right now?

(reaction:stop "Copland College" :SOUTH 0803)

Given a bus stop and a direction (southbound or northbound), tell me what buses I can expect to see. For example, in making the above query I should be given joyous news about the 315 and the 45, in chronological order.

How do I get from one stop to another in minimum time?

(reaction:trip "Copland College" "Woden Interchange" 0803)

This is the clever one: by tracking times and cross-referencing timetables, allowing for connections and alternatives, what's the best way to get from point A to point B? Arrange it by total travel/waiting time, with information about where you need to change over.

Where does this bus go?

(reaction:bus-stops 315 :SOUTH)

A plain list, as shown in the timetable leaflets: all the stops that a bus goes to.

Which buses go past this stop?

(reaction:stop-buses "Copland College" :SOUTH)

A plain list, as shown at most stops nowadays: all the buses that go past a particular stop, ordered by time.

The Lisp syntax is just for experimentation; my plan is to make this into a website, so that people can use it to achieve something without shuffling timetables.

This is made harder by the fact that, although every bus stop has its own unique identifying number, the numbers are only rarely displayed on the stops and never appear on the timetable leaflets. Also, not every stop is listed on the publicly-viewable timetables. What we really need is a complete dump of the entire route system, in XML format say, kept up to date and made available for people to do mashups.

I wonder who I need to talk to to make that happen...

Worksop, n.: A person who never actually gets round to doing anything because he spends all his time writing out lists headed 'Things to Do (Urgent)'.

- The Deeper Meaning Of Liff, Douglas Adams and John Lloyd.

Here's my list of projects I have on the go or off the boil. In no particular order, these are all the things I need or would like to get finished. They don't include tasks imposed from on high, either by my boss at work or my bossBeloved at home. My goal in listing them is twofold: to show people why they never see me any more, and to reassure myself that the list is not, in fact, of infinite length.

• The LaTeX/PHP data dump for Mr Death. This is one I'll actually be paid for. It's not urgent, because we can keep on producing the output files using a semi-manual method, but it would be much better to have it fully automated. For that to work, I need to finish adapting a program that was producing PDFs via a third-party library to produce LaTeX ".tex" files instead. It will involve some deep-down fiddling with TeX as well, I suspect, to get the layout looking as the user wants it, rather than as Dr Knuth decreed was best. The user is not as smart as Dr Knuth.
• Gratian, the editing program for Canon Lore. It has a number of hard-to-kill bugs that make it less than perfect, although it's really only a couple of thirty-hour debugging sessions away from perfect. (But then, by Cox's Law, so is every piece of software, so that's not saying much.)
• Instructions for Canon's minions. teffania Canon has found some volunteers, and my instructions have gotten them most of the way to full functionality, but there are still some holes to patch.
• Fix one or two more things in the Politarchopolis website and hand it on to a successor (I wonder if syridian could be coaxed...) because it's just not that much fun, really. Also need to do my quarterly report on that.
• Harkness, my "time agent" program (hence the name). The time-tracking program they use at work is arse pudding on toast, so I've been working in my spare time on an alternative, combining the least avoidable aspects of it with my venerable old time-keeping program Captain's Log, which blows it out of the water and no mistake. I can't work on it during work hours, ironically enough, because the official program is considered "good enough" by TPTB, so they can't authorise diverting me to replace it.
• A program to mirror my LJ onto my Wordpress blog. Since I don't trust the Russian Mafiosi in charge of LJ as far as I could spit them, I want to be able to turn my backups of my LJ into blog entries on my personal blog. To do that requires a program that will take the XML from the backups, massage it thoroughly and produce a different pile of XML that will slurp easily into WordPress. Given the lack of specifications and the ugliness of some of the data, that's a deceptively large job.
• General upgrades to all our computers. The Beloved and the EDoD are still on Ubuntu and Xubuntu 7.04, Feisty Fawn. I'm still on Kubuntu 7.10, Gutsy Gibbon, until I get confirmation that VMWare will work under 8.04, Hardy Heron, but I want to upgrade all of us to Kubuntu 8.04 to reduce pain, since it really does seem to be the best option.
• Clean up the Green Shed. Most of what's in there needs to be Garage Saled, desperately.
• Clean up the Beloved's study/sewing room. Ditto.
• Make curtains for the rest of the house. We've done the family room and the BatPup's bedroom. Tomorrow we do the kitchen. Between the Beloved's Op Shop-fu and the Spotlight voucher my Mum gave us for Christmas, we should be able to do the rest, although the sewing room will become a useful weapon in the fight.
• Finish the secondary chook run, which is a bit of a fox's breakfast at the moment.

That's about all I can think of at the moment, which isn't too bad. I may get to the end of the tunnel before the oncoming train arrives, if I can somehow manage to avoid spending evenings with my cuddly munchkins... Naaah.