Older blog entries for lloydwood (starting at number 9)

Added a last-filename-used indicator to fill up remaining window space; it's in SaVi development code. I should have thought of this a long time ago, especially considering that I'd quickly select a constellation and then guess what I'd loaded by looking at the coverage panel, before checking the orbital parameters.

The filename should probably be in the window titles, but I couldn't make that work in Tk. (Tk/window manager interaction seems quite poor and non-standard, if my efforts to generate window icons from Tk (which work in fvwm2, but not in KDE or Gnome) are anything to go by. This is probably really a failing of the separation imposed by the X model, and everything that flows from that.)

I'm forever amazed by the low usability standards that unix software reaches. But I think I'd prefer to address it by gradually improving the software rather than writing documentation excusing the behaviour by describing it. At least, that's my excuse for SaVi not having much in the way of documentation, much less properly-*roffed-out man pages.

18 Nov 2002 (updated 18 Nov 2002 at 22:14 UTC) »

SaVi now sports a -nologo option to its command line to allow unbranding. I'm such a fashion victim.

Interesting that The Economist says (in "Why Naomi Klein needs to grow up", 7 Nov 2002):

It takes journalistic skill of a high order to write page after page of engaging blather, so totally devoid of substance.

Would that the writers of the homogenized, here's-why-we-are-right what-the-rulers-of-the-world-would-want-to-hear The Economist had such skills. I decided I'd rather read careful observation, statistics and journalism than a magazine of op-ed masquerading as journalism, and my subscription to The Economist remains lapsed.

Draw your own conclusions. Always.

27 Oct 2002 (updated 27 Oct 2002 at 21:36 UTC) »

Thought about doing a SaVi release - this would be 1.2.2.

Pros: The Tk 8.4 compatibility needs to get out there (okay, so that's just a single added #if testing versions because they changed an api again), I can claim it's gcc-3.2-compatible (didn't have to do anything much new for that; in fact all 1.2 releases should build with gcc 3.2), and thanks to Matthias we're sort-of-Cygwin-friendlier, within the limits of the homegrown makefile. There are also improved standard file dialogs (Matthias again), a display bugfix that prevents clipping in the coverage tiling I did for 1.2.1, and some rejigging of the initialisation process. All of this stuff needs to be in a release.

Cons: Geomview isn't yet gcc-3.2 compatible (its C++ generates errors I don't yet understand), and Matthias' Cygwin suggestions for Geomview have only just gone into its cvs. So perhaps I should wait for a new Geomview release with Cygwin and gcc 3.2 compatibility before trumpeting my tiny bit of work on SaVi with a new release, given how the two are usually used together. I might be able to come up with the odd tiny bugfix meanwhile; the list of things to work on is quite long.

Anyway, an up-to-the-minute SaVi development tarball is available, so it's not as if the new work is particularly hard for anyone to get hold of.

In other news, my T-shirt design seems to be a minor hit. Dave started with 32 in stock, and 23 appear to have already been ordered since NTK announced the shirts on Friday. (I think it's going to become a well-worn phrase; it's taken up residence on my Nokia phone's LCD.)

Or maybe those shirts are just being bought because they've already run out of 'Corrupt Disc Inferior Audio' shirts.

27 Jun 2002 (updated 8 Apr 2004 at 20:56 UTC) »
SaVi and Tux - together at last! [Rats. The photo of Hannah holding a model penguin in front of the 9BB21 SaVi poster has been deleted.]

Good grief. Sourceforge tells me that downloads of SaVi have hit an all-time high (96.196% activity percentile last week, as opposed to 0.0% or <10%, whatever that means, since I've no idea what the distribution is - Zipf?), and I still haven't had to read any email about problems with the thing.

The increase in downloads could be down to the new introductory screenshot (showing Globalstar running under Geomview) on the webpages perhaps looking slightly more eye-candyish than the old introductory screenshot (showing Molnya running standalone). Or the increase could be down to me mentioning SaVi in these diary entries. Or it's due to creating a direct download link when updating pages for the new release. Or even Sourceforge improving their stats and graphing. (Still five significant figures? Hmmm.) I'd always wondered about some of those 0.0%. Hang on, if I don't believe the figures when they're low, why should I start believing them just because they're now high? And isn't talking about downloads and counters so 1995ish?

The lack of mail? Either it's working fine for everyone, or it's not working at all. Either way, since it's fully-disclaimed free software, I shouldn't be worrying about it. I just find it odd that I'm worrying about the fact that I shouldn't be worrying about it. And I find it odd that I find it odd. That worries me.

I must learn to let go; everyone else did.

Made yet another SaVi release, having decided that the release I'd released wasn't up to the standards that a release should be.

Not that anyone will notice. That's released code pretty much caught up with development; I can't see myself doing any more work on SaVi to satisfy the feature requests of its userbase this year, seeing how its userbase these days consists of me.

Impulsively did some more work improving SaVi while offline, using an ad-hoc local CVS tree I set up in a hurry.

Resulting SaVi screenshots look pretty, but now I have to figure out how best to merge that local CVS tree back into the master SaVi Sourceforge tree.

Of course, my local cvs tree has overwritten all tags, so I can't simply commit. CVS convenience comes with a concomitant cost. Aiiiieeeeee!

I'm still hacking, on-and-off, on SaVi as time permits.

Even though, as far as I can tell, no-one else on the planet is currently using it enough to find things worth complaining to me about that I could try and fix. Or at least, things that aren't already listed as being known about in the copious bugs/wishlist textfile included in the tarball.

And even though, at this point, there can be no possible future value to me in ever adding further value to SaVi. It's not as if I'll ever use it in the day jo^H^Hcareer, after all.

Frankly, I'd have expected my behaviour to be more rational by now.

I suspect that Advogato could, eventually, turn out to be the first true weblog software for a focused community.

'Recent diary entries' is, after all, simply a very fast way to read a number of aggregated weblogs without all of that tedious clicking that takes you nowhere much. But without frilly themes and customization, Advogato will never be for the masses.

The trust metric stuff might be a red herring; I can't say that I don't trust e.g. this guy; it would not be possible to explicitly distance him from myself in a graph we were both connected to.

I don't think trust is transitive, commutative, or associative.

15 Jan 2002 (updated 15 Jan 2002 at 16:17 UTC) »

Created Advogato account. Looked around. Edited entry.

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!