wingolog

mentalities

Yesterday, biking the same road I bike every day between the office and home, I caught a mental whiff of past places, of cliffs and sea, of leaves and mountains. Somehow these mentalities are very far away now.

bad ideas

You'd think people would realize by now that naming your project "new foo" is a bad idea. All bits rot with time. What is new either becomes an historical curiosity or is forgotten.

So software developers, when it comes time to name your Frob project and the idea Frob-NG comes to mind, chuckle, make a joke about star trek and keep looking for a different name. You can thank me later.

the practical arts

Sergey writes desiring a way to prevent the military from using his software. I probably agree with his perspective on the military. I personally think that our modern armies are regressive implements of illegitimate state power, the US foremost amongst them. But the military's relationship with science and technology is a bit more complicated than that. The first computer's job was to calculate ballistics trajectories for the US army. The current top machine's job is to simulate hydrogen bomb explosions. The same technologies enable the internet and gene sequencing.

Technology is a tool, in the end; as a tool-maker you have to decide whether it's worth it to do what you do. Personally I wouldn't work at Los Alamos even if I were just modeling the weather, as I feel that is too close to the bomb. But I might take the military's money to do other development (and holy jesus do they have a lot of money -- a large piece of the pie).

Everyone draws their own lines. Putting your software under a different license is one way to do that. It's probably not the most effective fight, though -- with that much money in their hands it doesn't really matter what you do. Better to get political and fight at the source, if that's your thing.

work

It's been brought to my attention that I haven't talked about work in a long time. So here is what is up in the Fluendo basement.

Flumotion-the-streaming-server is what I hack, that and some other software for Flumotion-the-streaming-company. We're selling services on a platform that uses Flumotion to stream live and on-demand video and audio, with some extra layers for load balancing, transcoding, proprietary formats, logging and billing integration, and the like. Of course, the streaming software itself is Free.

We're just wrapping up a development cycle on Flumotion and the platform software now. I think that Flumotion users will find that it's much more robust, especially in the face of transient network issues, CPU usage spikes, and in useful feedback to the person administering the whole thing. There are a few new features as well -- recording to disk based on an ical file, a static file server, serving multiple mount points on one port, and a few more. But to me the important thing is that the whole system is less fragile than it used to be, which should be a more positive experience.

The downside of all of this for me is that to turn 0.3.2 into 0.4.0 I have to do a lot of testing, which is tedium. I dislike this so much that I wrote a graphical forkbomb to run the tests instead of sitting down and running the tests. Maybe it was a nice time investment? Unclear. Anyway, I believe I've moved out of Flumotion testing and am now testing $232,020 worth of proprietary platform software, according to sloccount. I would have to say that I prefer making the bugs than testing for them.