3 May
2004
3 May
'04
9:27 a.m.
On Mon, May 03, 2004 at 12:50:56 +0200, Tim Goetze wrote:
[Steve Harris]
Its disposable in the very little investment was made in implementing the
core, it was just a bit of headscratiching and a pretty small .h. If its
has a life beyond then introduction of GMPI then its obviously OK, and
there was a need for it. Theres no problem with it living on, it just
seemed silly to invest any significant ammount of effort in designing and
implemented something more fully featured when GMPI is "just around the
corner".
On Sat, May 01, 2004 at 03:33:14 +0200, Tim Goetze
wrote:
i like the spec as it is, but i'm not entirely comfortable with the
'disposable' attribute.
for one thing, i'd not like to see the idea disposed of before it
becomes used ... :) i'm inclined to think that free-form docs are
sufficient. it would be
nice to have a complete conf language but the effort implementing that
is likely to be huge or require yet another library or both, and maybe
not even worth it.
Yes, The D stnads for disposable - its just a quick hack to get embedable
softsynths without building much thats new - it reuses existing
technologies as much as possible - the core DSSI code/spec is tiny. so far, the only weakness i see is the lack of support
for GUI-less
operation, which can easily be amended by making conf documentation
mandatory (synthesis parameters that aren't accessed through LADSPA
ports should also be documented i think). sure, you can reconstruct
that info by going over source code, but i'd rather you needn't.
We dont yet know how many synths will use configure() I suspect not that
many, and if they dont need that then host built UIs will work perfectly
well.
I do see your point about the documentation recommendation though - I have
no problem with it myself, but I'm not sure how useful it will be.
- Steve