Older blog entries for Radagast (starting at number 8)

I've done it again. It's 3:40 AM, and I've been awake since 6 PM a day and a half ago. That's 34 hours. The hallucinations are creeping up again. I think I'll go and sleep now.

The Peer Press mailing list (which I stupidly used the wrong URL for earlier) seems to be picking up speed. Now to convince people that HTTP is not a magic bullet, and get working on the prelim spec for the DTD.

Hans Petter is still tied up in marginally revenue-generating work, but at least there's light at the end of the tunnel. Specifically, the Monday end of the tunnel, which is when we're supposed to deliver. I might be crazy, but it looks like it's on track for then. He's been toying with some cool genetic algorithm stuff on his breaks, I notice. Could also be very cool.

Paid the advance on our GUADEC tickets today. We're going to be in Norway from the 20th of March to the 10th of April too. If any Norwegian free software hackers want to meet (preferably in Oslo), tell us. We'll also be arriving in Paris on the 15th, so there's a spare night where we'll be doing nothing formalized. Any others going to be there that night?

My God, it's full of stars. Or wait, that's sleep dep again.

It seems I wasn't the only one who thought it was possible to do better things than Slashdot. The article I posted got lots of great response, and today I set up a mailing list for discussing where to go next. I have no doubt something like this could make Slashdot largely irrelevant. The interesting question is, could it be an arbiter of change in the larger frame of things, like traditional press? I've thought about this a lot, and the scary bit is, I can't see why it couldn't be. Not seeing any faults is usually a warning sign that you're not looking hard enough, but I don't know. We'll see what happens.

Work rolls on. Our hunt for investment is coming to an end soon, I think, and with the outcome we wanted. Nothing certain yet, of course, but I'm hopeful.

I'm starting to really look forward to GUADEC. I feel like I know so many of these people, now I need to meet them in the flesh. I suspect much late-night discussion will happen.

Reading the /. thread attached to the GNOME development roadmap article, I realize the reason I bother is the same reason people watch autopsy photos, or The Jerry Springer Show; sick fascination. And to think that back in 1997/98, I had a Slashdot habit that required me to check it at least 20 times a day...

Iain checked in some Loom code, it seems. I haven't had time to try to compile it yet, to see what it does, but I'm happy stuff is happening there.

Hans Petter seems to have gotten the Oracle crap under control, and Peter Murphy was most excellent last night. I didn't even know Mexico City had a goth community.

Read Raph's diary entry, in particular the part about commercializing web sites, and the principles behind Advogato, with great interest. My question to all this is, is it impossible to be "yet another IPO-track dot.com website" and still have great community, and not sell your soul? I think it is. Slashdot, for instance, didn't change much when they were commercialized. The overexposure killed them, because they didn't scale, but I think having money didn't make much of a difference in content. Maybe it made too little of a difference, since it seems they're all happy sitting on their ass with their IPO money now, and just post articles now and then. I'm getting more and more convinced something new is in order.

Nice Saturday. Hans Petter is working on finishing the project Conglomerate sprung out from. He's wading through literally thousands of pages of Oracle documentation to get the storage part working. After all I've heard about the wide-spread use of Oracle for "enterprise" applications, I had expected it to be well-designed, homogenous, and thoroughly logical. But it seems that it's more like the opposite, an incredible collection of hacks and old cruft. According to Hans Petter, it's even more dirty than the code to come out of the worst, most evolutionary free software projects (I assume this brings it even below the level of Perl, and that's a scary thought.

Python bindings to Flux are coming along nicely now. I got a boost since some people contacted me and actually wanted to use it. It's in CVS, and I aim to finish them (well, reasonably finished) for the 0.5.0 release, which should be in about two weeks.

Been thinking a lot about web sites and content lately. There's a lot of crap out there, when it comes to free software sites. Slashdot has gone the way of Usenet long ago, of course, Freshmeat pretty much always sucked, and the whole VA/Andover juggernaut seems to be quite happy with where they are, and the IPO millions are nice to have, thank you. Maybe it's time for something new. Got some ideas, talking to the people here at work to get a go-ahead for that project. We'll see.

The recent controversy over certification, mostly a whine-fest by the Jabber people, shows us more than anything what the results of Eric Raymond's rhetoric truly are. People think that because they throw around "Open Source" and other buzzwords, their project is significant. Well, it's an instant messaging system, whoop! No matter how many buzzwords you use, like "streaming XML", "no chunks" (which is utter bullshit), and whatnot, it's just a workalike for ICQ, with perhaps some extra features. Annoyance.

Going to see Peter Murphy tonight, should be good.

On my 36th waking hour now, for reasons I'm not entirely clear on myself. It started with me falling back to my old work as a designer to cover for a large amount of work at one time, and it kind of got out of hand. Still, I've made all the design for two web sites in this time, and according to both me and others, they're some of the nicest I've done. Now to implement the HTML, but that'll happen after I sleep.

Flux 0.4.1 was released, on the heels of 0.4.0. With this new release comes new web pages, a lot of very nice new functionality, and lots of optimizations. Hans Petter seems happy and worked up over it. I just wish I'd had time to work on the Python extensions lately, but I haven't. I'm scheduling that for the end of the coming week, I think.

Also, Loom is kind of stagnant, mostly my fault, Iain has actually done work on it, I haven't. Need to look at that, preferably at least poke around in it this weekend, see what happens. I took a look at The Sims today, the new game from Maxis, and that looks absolutely fantastic. I know I need to get it, but I also know my productivity will suffer terribly. Good thing is, it's iso 3D. I'd really like to clone this when Loom is more or less working, but it'll be a lot of work. Better not put more items on my todo just yet. Hans Petter says he wants to port our old favourite networked strategy game, Empire, to Loom when it's more or less working, though, so this could get fun fast.

Went to see Bringing out the dead today. Brilliant movie. Scorsese is a genius. Same day as Hans Petter and I discovered we both love his movie After Hours. Weird coincidences.

Sleep now, I think. My vision is getting slighly blurry, and I have a headache.

Long time since I wrote anything here. Lots of things have been happening, i comitted the new GNOME webpage layout, people seem happy with it. Flux has a lot of neat new features now, like fully crypted sessions. We're toying with prototypes for a VPN product based on this. Hans Petter is working on object database functionality for Flux now, and it's looking really nice. This is going to be a big boon both for Conglomerate and FIDEL.

I've more or less mapped out what I'm going to be working on in the next six months, and it's looking quite interesting. The Python module for Flux is going well, in addition I'm hacking on an admin console for this VPN thing, and later I'm going to be doing interface work for FIDEL, and also starting up my toy project of writing a generic isometric 3D engine for games using the GNOME canvas. Strategy games are fun, and so is the canvas.

Also went investor-hunting the last few weeks. Seems a lot of people are interested in what we're doing, but there's a bit of a leap from there to actually giving us the money. We'll see.

We didn't know what we started. Since we announced Conglomerate a couple of days ago, the response has been overwhelming. Lots of mail, lots of encouragement, and general inspiration and ego-stroking. It seems Hans Petter and I are going to GUADEC as more or less a direct result of this. Now all we need to do is juggle all the damn projects, get a test release of Conglomerate within a week, FIDEL within new year, and then we plan to do the new Conglomerate parser in January, alongside the reimplementation of the main editor display as a Canvas.

The fun never ends. Thanks to everyone who's said good things about our stuff lately.

So, we finally got around to announcing Conglomerate today, on Freshmeat and Gnome news. We thought it was about time, since there's been increasing interest about the project, and after talking to Miguel, we got some new ideas and input, so we decided that now is the time. I produced a record amount of content for the webpages, and fixed the design, so everything should be set. I hope people will get on the mailinglists and start giving their opinion.

Hans Petter got Daniel to check out the 0.3.0-pre of Flux from CVS today, so the textmode client for FIDEL is finally on its way. December 4th just started, slightly below 28 days to get a working FIDEL. I still think it should be possible. But then again, I might be mad.

Interesting day, got a tip from Ole Aamot to tune in to #gnapster, and got there in time to watch the end of the fireworks, as well as some excellent diplomacy by Nat. Bad part: I resurrected my IRC habit. This might be tough to shake again.

On the professional side, almost finished up the client web site I'm coding on, so I can turn to more interesting stuff soon. Crimz is wrapping up documentation for flux, his neat networking, data storage and general utility library, so it seems coding on FIDEL will commence this week. We have one month to get it working before Y2K eats FIX' current BBS software. Should be possible. Also got a chance to talk to Raph about trust metrics, and it's fairly certain that a trust metric system is going into FIDEL at some point. The number of users and amount of activity on FIX is already too large to be easily administrated by hand.

Oh, well, back to coding.