OT: Re: [LAD] Source code for Open Sound System now available under GPLv2 and CDDL 1.0

pete shorthose zenadsl6252 at zen.co.uk
Fri Jun 15 00:23:41 UTC 2007


On Fri, 2007-06-15 at 00:01 +0300, Hannu Savolainen wrote:
Pedro Lopez-Cabanillas wrote:
> > On Thursday, 14 June 2007 00:41, pete shorthose wrote:
> >   
> >> On Thu, 2007-06-14 at 00:42 +0300, Hannu Savolainen wrote:
> >>     
> >>> Hi folks,
> >>>
> >>> Now it has happened. Open Sound System is finally open sourced.
> >>>       
> >> http://4front-tech.com/hannublog/?p=5
> >>     
> >
> > Quote1:
> > 	"However the ALSA API is still almost completely undocumented 
> > 	(after three years of it’s release) so how can anybody expect 
> > 	that programmers could develop good applications based on it."
> >
> > Quote2:
> > 	"For the above reasons the benefits of OSS are widely ignored:
> > 	[...]
> > 	* It’s fully documented (OTOH some parts of the documentation 
> > 	are still under construction)."
> >
> > Yeah. The ALSA API is also fully documented, but some parts are
still in a 
> > work in progress state...
> >   
> Interesting. Is there some document that explains how to use all the 
> 1500+ functions exported by alsa-lib? I know only the doxygen source 
> listing but I would not call this API documentation.

please don't use disingenuous arguments around here. these lists are
populated with double smart bastards and any number of us are capable
of seeing through such misdirection. 

you said alsa was "not documented". clearly this is both a gratuitous
overstatement and factually inaccurate. particularly so in light of 
your claim that by contrast, oss is completely documented (albeit
followed by the contradictory claim that parts of it are under
construction?!) to then defend yourself with an expectation of 
100% documentation (you said "*all* the 1500+ functions") betrays
the disingenuous slight of hand that is the very signature of your
comments on your blog. just because something is not fully documented
does not mean it is not documented.

to be optimistic about your own projects and pessimistic about those
you perceive as competitors is perhaps understandable, but non the
less bogus for it. alsa has many faults and an honest appraisal of
the benefits and drawbacks of oss vs alsa would go much further
in convincing the majority of us who are immune (if not allergic)
to PR than a partisan rant replete with a raft of misdirection, 
straw man arguments and outright invention. 
(i mean seriously, to infer that alsa requires a realtime kernel
(blatantly not true) or that you must use callbacks, (ignoring
the conventional push style read/write interface...) or that
alsa requires libjack.. or spuriously associating callbacks with
goto and come-from statements.. the list goes on and on.
(but ho hum, it's unfair of me to trawl through
the entire blog detailing every contended point so i won't)

if you run around unfairly shitting on alsa then you stand a very
good change of burning through any goodwill that exists as a result
of you releasing oss under the GPL (among other licences). 
the alsa developers stepped up to the plate at a time when you
did not, and i for one am very grateful that they did.  

> However the above blog posting is not necessarily valid any more. 

a belated admission coming only after redoubling your assault 
above. the literary equivalent of apologising to someone *before*
you slap them. it's only effective if those witnessing aren't paying 
attention.

> Takashi has released the salsa library after I wrote it. 

as i understand it, salsa is ostensibly intended for embedded use. 

> It's certainly 
> good development if all the dead code could be removed from the ALSA 
> library API by replacing it with a new one that has just the
functions 
> that are actually necessary (used by more than one application).

can you provide any statistics pertaining to api usage in the wild or
or is this just speculation on your behalf?
while i'll freely admit that bloat is a common enough complaint with
regard to alsa lib that i wouldn't go out of my way to contend it 
generally, whether or not oss provides sufficient granularity to be
useful for generic and specialist (audio production) uses cases alike
remains to be seen. 
bragging about the simplicity of your api without providing much by
the way of example doesn't impress me much. quite the reverse in fact.
again, you are doing yourself and oss no favours by aggressively 
attacking alsa. if oss is, on balance, superior then that is sufficient
to convince me.

 
> Now it 
> looks like that no known applications are using majority of the stuff 
> exported by alsa-lib.

i could sweep the taj mahal with that statement. no attempt to 
substantiate that at all? 
the sad thing is, it might even be true, but with purely speculative
delivery like that, you are simply poisoning your own well.

> 
> > What happened to MIDI/sequencer?
> > http://4front-tech.com/hannublog/?p=6
> >
> > Quote3:
> > 	"The real problems started few months after that when I tried to 
> > 	document the sequencer/music interface.  [...] Due to lack of 
> > 	time I gave up but promised to write the documentation later."
> >
> > Quote4:
> > 	"The bad news is that there is no MIDI support in OSS 4.0. There 
> > 	are few bugs in the MIDI code and we decided to ship OSS 4.0 
> > 	without it [...] MIDI support will be included in OSS 4.1 
> > 	(hopefully within this year)."
> >
> > OK, lets talk about MIDI after some months.
> >   
> Let's do so.

i second that. and hopefully with more fact and less fat.
so anyway, i wrote this response in the tone of Hannu Savolainen.
how does it feel? angry yet?
perhaps you see my point. 

what ever the motivation for the switch to include GPL licencing,
i appreciate your efforts. i'm neither anti oss nor pro alsa
but i am a horribly conceited person who doesn't much appreciate
being taken for a mug. and just because there's neither a hierarchical
management structure nor meaningful financial incentives to work on
linux audio (for the most part) it doesn't mean that the metrics of
developer loyalty to any given system simply fall through to 
technical superiority. you yourself have alluded to this already,
although contrary to your beliefs in a grand alsa conspiracy, in 
my experience, i've found that people react against any perceived
dishonesty or expedient manipulation of the facts, rather than
falling for it.

i'm sorry if you feel i'm being unduly harsh,
but then, if you can dish it out...

pete.




More information about the Linux-audio-dev mailing list