Subject: RE: [linux-audio-dev] Still I cannot understand why...
From: Ivica Bukvic (ico_AT_fuse.net)
Date: Sun Dec 16 2001 - 05:53:19 EET
> Please, no more half-way compromises with OSS... kill it, do
> not even consider "backwards compatibility" for old OSS apps!
> Are there any worthwhile salavaging in light of newer developments ?
> Every ounce of effort towards maintaining (crappy) backwards
> compatibility for old OSS mainly/only systems takes away effort
> from enhancing true ALSA apps.
The reason why I am so hung-up on backwards compatibility is because
this is one of the unique traits of linux. Both winhose and especially
crapple, change their architecture every five minutes both for profit
sakes and because their architecture inherently sucks. On top of that,
when an OS exemplifies in an extreme backwards-compatibility streak,
that means that that OS has a good foundation and as such instills
consumer's confidence. Now, I am not saying that the OSS is a good
foundation, but at least having a small part of the code that could not
be addressable by the new programming endeavors, but rather to keep the
backwards compatibility, would make sure that the existing
apps/commercial software runs and runs good (i.e. Loki games, and older
audio stuff that might not be developed any further, yet is extremely
good, such as RTMix). I think we would want to be inclusive, not
exclusive, and if the daemon was built properly this part of the OSS
backwards compatibility would not be even noticed or possible to address
in the new apps (if the docs were carefully implemented), and as such
would not pose a threat of prolonging OSS's sloppy reign over linux
audio.
You also have to keep in mind that here we are talking about burying
80+% of one commercial company's sales (4fronttech, author of the OSS,
needless the mention other commercial companies who depend on OSS, such
as Loki games), and as such it will be extremely hard to do, without an
extreme opposition from the company who has the bucks to spend in order
to counter our move with propaganda. Thus, I think we need to do this
transition as smoothly as possible, not stepping on anyone's toes, so
that we encounter the least possible amount of friction along the way
and make the transition as fast as possible.
> That's part of the deal with open source and personal development. Who
> wants to really work on someone else's project when you can do yer own
!
And that is exactly what makes us weak. Everyone re-inventing their own
version of wheel will ensure that we all stay in the stone age and not
move even one inch from that point on the evolution scale... Why don't
we make it "everyone's project" then :-) ?
> Anything strictly ALSA based (particularly so newbies don't get
confused
> by two seperate code paths) and the smaller and finishable the better,
> lot's of people have big grand plans for larger projects but getting
> usable (good demonstration) apps in a timely manner would also be
> beneficial for a lot of LAD fringe-dwellers.
Again, if the daemon code and docs were implemented carefully, oss apps
would be extremely discouraged, if not impossible.
This archive was generated by hypermail 2b28 : Sun Dec 16 2001 - 05:48:18 EET