[linux-audio-user] Common linux audio layer

omjn looplog at looplog.org
Sat Jan 8 03:01:30 EST 2005


Lee Revell wrote:

>Hmm, some people seem to have the impression that Linux audio is the
>byzantine beast that it is because Linux users like it that way.  This
>could not be farther from the truth.  Many of us are working feverishly
>to make it "Just Work" the way Mac and Windows do.
>
>  
>
hi Lee,

If by "some people", you mean myself, then your impression of my 
impression couldn't be farther from the truth!  I am quite aware that 
linux audio, and linux in general, is a product not just of users 
desires, but of the politics associated with many differing projects 
working in parallel, each with different end results in mind.  To put it 
another way, the complexity and diversity us linux seems a direct result 
of its openness. I congratulate any and all working towards a goal of a 
'just works' linux audio system.  I was merely trying to add a pragmatic 
voice to the discussion - commercial R&D is often driven by top-down 
management style directives made by those in charge.  This is a 
deliberate structure I would say that is set out to ensure that 
duplicity doesn't take place in the development process and to ensure 
development resources are efficiently utilised.  Without this kind of 
streamlining it is difficult to implement top-down directives in the 
first place (such as "all linux distros should use this common linux 
audio layer because we said so").  I'm sure Linus can get away with it - 
he invented the thing in the first place.

In linux, so far as I can tell, this top-down approach is not so 
evident, as there are many many tops, each with its own agenda.  This 
makes the kind of cordination that is required for unified system 
architecture development a lot more difficult I would have thought.  Who 
would you propose is responsible for making the decisions as to what 
should be included and what shouldn't?  Where does the managerial 
authority lie to judge one audio layer as the right one?  Most often, 
there seems to be a vague consensus amongst a few like minded developers 
and that is taken as the direction for a particular aspect of 
development.  If anyone is serious about user-centric design, where are 
the usability studies?  Where are the surveys of current users to find 
out what is working for them and what isn't?  Certainly, there are 
testimonials on this list and elsewhere, but I would be reticent to take 
the opinions of a few english speaking members of the linux audio users 
email list as representative of all Linux Audio users opinions.  I've 
often heard it said that the direction of linux is directed by those who 
shout the loudest having declared themselves spokespeople for the 
average user.  That seems fairly ad hoc, and I doubt it would produce 
results that suit the fairly large demographic that this undefined 
common user would make up.

So again, I'm not dissing any attempt at making things easier.  As I 
stated, I think Jack is actually one of the best architectures for audio 
on all systems, at least conceptually if not in implementation (though 
it seems pretty damn close to being there).   It gets my vote by far for 
the best approach to unifying the audio systems out there. 

I just get a little angsty when I hear people proclaim they are doing 
things for the user, especially when they reduce one user's carefully 
worded argument down to a two sentence impression ; P.  Are you really 
championing the users cause, or are you just doing it to make it easier 
for yourself?  If that is the case say so, don't be afraid to be 
honest!  Market speak floods enough arenas of life...  If there really 
is this groundswell of user desire for a unified or common linux audio 
layer, then fair enough, document it and I'll shut my peephole and go on 
being a stick in the mud for developmental diversity.  Who knows, that 
document might then serve as suitable evidence for moving to one audio 
layer/server over another and could even induce some of the coordination 
between developpers that would be needed to make it reality.  I just 
think you would need to be fairly comprehensive about putting such a 
document together.  I would volunteer, but I'm in the last stages of 
writing up my PhD so I don't have alot of time on my hands.  I'd 
certainly help in any way I could if people think it helpful.

As an end note I would question the presumption that mass-market 
adoption of linux as a unified desktop environment is a good thing.  
That may or may not be the case, but as far as I can tell, there have 
been very few mass market adoptions of products not developped using 
mass production techniques (ie managers in plush offices telling worker 
drones to implement their market researched visions by employing 
mind-numbingly repetitive yet highly efficient development or production 
techniques).  If you want to make a mass market product, this is usually 
what it takes.  As far as I can tell, Linux development is not 
structured this way because many developers do it for fun. 

Anyway, you call it a behemoth, I still like the bazaar metaphor...  
When I say I like linux the way it is, I mean that the development model 
is not so explicitly driven by mass-market rationales, not necessarily 
that I consider it feature perfect.  If there can be a common layer 
implemented without dictatorial directives, then I'm all for it 
(especially if it's based on Jack.  Did I mention that I really like 
Jack?). 

best regards

Michael

ps- I know my writing style can be a little aggressive, and for that I 
apologise.  Damn those years of critical theory...



More information about the linux-audio-user mailing list