Subject: [linux-audio-dev] Re: [linux-audio-user] Audio routing issues for linux..
From: Juan Linietsky (coding_AT_reduz.com.ar)
Date: Mon Jun 10 2002 - 10:43:13 EEST
On Mon, 10 Jun 2002 14:37:42 +0900
Patrick Shirkey <pshirkey_AT_boosthardware.com> wrote:
> Juan Linietsky wrote:
>
> > Probably the easier and more natural approach to this is just
> > integrating JACK to ALSA in some way.
> >
> > What do you think?
> >
>
> Your idea has been discussed in length by Abramo and Paul last year
> and although Abramo was adament that it should be in the alsa-lib
> that this is done he was sorely contested by Paul who used all the
> other pro/closed source versions of JACK as examples why it should
> not be included in the alsa-libs.
>
You have no idea how much I regret not being in the list last year,
as i would have sided wit Abramo :)
> Mostly because to do it correctly he argued that it has to be based
> on a callback design and that is not part of the alsa-lib. Hence
> JACK.
>
Sorry, obligatory question, what's wrong with blocking calls? :)
To me callbacks is more like the idiocy of the winapi design.
Most windows programmers end up transforming the callbacks into
polling functions anyway.
> Richard Furst believed that it was not necessary to have a callback
> design and he coded up LADMEA. Abramo was happily integrating alot
> of the ideas for Jack into ALSA but then he lost his funding from
> Suse and has got another job which takes up most of his time and
> energy.
Ah, I'm sad to hear that.
>
> The general consensus is if you don't want low latency then use
> ARTS/ESD if you do want low latency use JACK.
I think, in a sense, It's sad that most of us are programmers.
I find the "use the ARTS/ESD-JACK for whathver you need"
definitively not either an userfriendly or programmer friendly
approach. Why not learning from Takashi's approach with the sequencer
api? It could have been a separate lib too, but it wouldnt have become
so useful or popular. It also _rocks_ that even OSS programs can use
it via the emulation layer.
>
> JACK is supposed to be faster to integrate than ALSA once the
> callback principal is part of the lowlevel design of the app.So
> instead of spending time on adding ALSA support get everyone to
> support JACK instead and then they don't need to spend time on ALSA.
>
Not to be pessimistic, but I think such thing is not going to happen.
Also I find the whole idea redundant. If JACK is easier and faster to
integrate then it should replace the ALSA api on that matter.
Developers, and specialy new ones, will allways support the official
api FIRST.
Juan Linietsky.
This archive was generated by hypermail 2b28 : Mon Jun 10 2002 - 10:40:31 EEST