From: Guillaume L. <gla...@te...> - 2002-12-27 22:11:10
|
On Friday 27 December 2002 19:26, Chris Cannam wrote: > I should say that I'm not (at the moment) about to get too deep into > any question of whether this experimental version is doing "the right > thing", that is, I'm not getting involved in whether it's going to > continue to work with repeats or audio (until I've tested them) It wouldn't anyway, there's no support for either of these. > In other words, being as good as what we have now but just different > isn't good enough. > > Okay, so you knew all that. Of course. To clarify things, I'm not advocating for this code to replace the current scheme. Even though right now it's much simpler than what we have, it's lacking too many capabilities (metronome, sysex events, repeats, audio) for that to be any relevant. As I said, this is a proof of concept, actually not even that because I was expecting it to be unmoved by any edit operation, and it seems it's not (which I'd really like to understand why). I'm not too disappointed because it seems to still achieve farily good results given that I've worked on this for just a bit less than a week, but right now I can only consider this as an experiment. I'll keep trying improving it for a few days, but if it still doesn't offer obviously better performance than the current scheme, then I'll just shelve it. The code will still be in CVS in case we want to re-examine it later. -- Guillaume. http://www.telegraph-road.org |