We are definitely going to release it at some point... There has been a pretty big response from Slashdot folks. I've gotten quite a bit of really encouraging emails. At this point it's just a function of time (since this isn't my primary job). But there is clearly interest in seeing the app get released.
That's a pretty good idea with the looking-for-a-hit -after-the-GI approach. You probably have to toss in there that if another GI happens first within the window of time (before a hit), then you switch around and expect a hit from the other side. Assuming it's possible to detect the GI sound, I bet that would be pretty successful.
If you've ever released an open source app, you know you can't just toss it out there. You have to package it, document it (at least enough to get people going), and support it. None of these are particularly enjoyable. Especially for an app like this -- video analysis is really finicky (at least mine is:) ). In fact, when I brought it to work after working on it at home, it just didn't work. It took me a good 45 minutes of screwing around to realize that somehow the color calibration got messed up on the card (the red was pushed way up).
Secondly, the whole point of that page (if you read to the end) was to see if there actually was any interest in such a beast. We love it here, but it's sort of a dorky niche product. I wasn't going to go through the pain of all the things listed in the first paragraph if there was one other dude in the world that was even mildly interested in it.
And lastly, but perhaps most importantly, is that I used some of the libraries from work to speed up development, which means I have to rewrite those parts of the library before I could do any sort of release.
By the way, the low level video capture interface is released under LGPL at http://jmds.dev.java.net.
Oh yeah -- One last thing. That someone would spend three weeks of free time developing an app to track stats for a videogame is pretty sad. That someone would instead spend their free time making up a story about spending three weeks of free time developing an app to track stats for a videogame is REALLY sad. I like to think I qualify for the first sad and not the second.
We've actually talked about adding in audio detection also because some events (GI's in particular) have really distinct sounds though don't necessarily have a very distinct video event (though I think GI's do actually flash for a few frames, which might be detectable). But audio detection seems like it would be a lot harder to do, so I just haven't done anything with it yet.
Slashdot Mirror
Good point :)
We are definitely going to release it at some point ... There has been a pretty big response from Slashdot folks. I've gotten quite a bit of really encouraging emails. At this point it's just a function of time (since this isn't my primary job). But there is clearly interest in seeing the app get released.
That's a pretty good idea with the looking-for-a-hit -after-the-GI approach. You probably have to toss in there that if another GI happens first within the window of time (before a hit), then you switch around and expect a hit from the other side. Assuming it's possible to detect the GI sound, I bet that would be pretty successful.
If you've ever released an open source app, you know you can't just toss it out there. You have to package it, document it (at least enough to get people going), and support it. None of these are particularly enjoyable. Especially for an app like this -- video analysis is really finicky (at least mine is :) ). In fact, when I brought it to work after working on it at home, it just didn't work. It took me a good 45 minutes of screwing around to realize that somehow the color calibration got messed up on the card (the red was pushed way up).
.
Secondly, the whole point of that page (if you read to the end) was to see if there actually was any interest in such a beast. We love it here, but it's sort of a dorky niche product. I wasn't going to go through the pain of all the things listed in the first paragraph if there was one other dude in the world that was even mildly interested in it.
And lastly, but perhaps most importantly, is that I used some of the libraries from work to speed up development, which means I have to rewrite those parts of the library before I could do any sort of release.
By the way, the low level video capture interface is released under LGPL at http://jmds.dev.java.net
Oh yeah -- One last thing. That someone would spend three weeks of free time developing an app to track stats for a videogame is pretty sad. That someone would instead spend their free time making up a story about spending three weeks of free time developing an app to track stats for a videogame is REALLY sad. I like to think I qualify for the first sad and not the second.
- Mike
We've actually talked about adding in audio detection also because some events (GI's in particular) have really distinct sounds though don't necessarily have a very distinct video event (though I think GI's do actually flash for a few frames, which might be detectable). But audio detection seems like it would be a lot harder to do, so I just haven't done anything with it yet.
That really is a much cooler name :(