Re: future ALSA development

[Takashi Iwai <tiwai@suse.de>]

At Mon, 23 Jun 2003 14:01:08 +0200 (CEST),
Jaroslav wrote:
> 
> > > * investigate a lisp integration to the current configuration syntax
> > >   - we need to describe the relations between high level abstract
> > >     layer (ordinary mixer) and current universal controls (very lowlevel);
> > >     it seems that the simple configuration is not able to describe
> > >     these (in most cases) very complicated paths
> > 
> > yep.
> > 
> > >   - note that describing of these relations might be used also for
> > >     another mixer interfaces (simple mixer for example)
> > >   - I don't rely on lisp, but what another interpreter with function
> > >     definition has only 22kB binary (slisp-1.2 - i686)?
> > 
> > surely it's nice but i don't think it's good to introduce more
> > complexity into the asoundrc syntax.  or, if you mean an external
> 
> We need to introduce the configuration extensions. At this stage,
> I want to investigate, if the runtime configuration modification (runtime 
> functions) might be converted to lisp rather than doing more or less ugly 
> hacks in our configuration syntax parsers.
 
my afraid is that the more control flows like conditionals may lead to
a difficulty for a parser utility.  i.e. the parser itself would be
like an interprerter.  but it's true that an extesion is needed
anyway...

> > database for describing the mixer configuration, it would be the
> > outside of the (old) simple-mixer API, IMO.
> > 
> > in that case, anyway, we can drop simple-mixer API and replace with
> > the new one.  the strategy used in the simple-mixer API can be
> > integrated into the new (ordinary) mixer API.
> 
> I think that the ordinary mixer API and simple mixer API does not have 
> much common parts for complicated hardware. The ordinary mixer API is very 
> highlevel and tied up to specified playback / capture streams, but the 
> simple mixer API is still only a translation between universal controls
> and mixer applications. I think that both should be configurable, but the 
> question is how much of these configurations can be shared. I would keep 
> the configurations separate.

IMO, the current simple-mixer API is still too complicated and not a
good abstraction.  as you mentioned, there is a big gap between the
low-level expressions (control API) and the reasonable mixer
appearance.  the extra information for the mixer would be needed to
fill this gap, too.


> > about lisp: i don't have objections.  i like lisp, too :)
> > the parser can be quite small, so if we build the parser in alsa-lib,
> > it's a good choice.
> > 
> > other people might propose XML.  then it becomes to a question whether
> > alsa-lib should rely on other libs...
> 
> I think that XML is too overkill for our purposes.

i don't think it's overkill but just a question of library
dependency.  and, writing a (good-readable) XML by hand is (not always
but often) a hard work.  so i myself don't buy this unless any other
good reason appears.


Takashi


-------------------------------------------------------
This SF.Net email is sponsored by: INetU
Attention Web Developers & Consultants: Become An INetU Hosting Partner.
Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission!
INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php