Re: [Jack-Devel] JACK on mobile

On Thu, October 17, 2013 1:40 pm, Paul Davis wrote:
> On Wed, Oct 16, 2013 at 10:13 PM, Patrick Shirkey <
> [hidden]> wrote:
>
>>
>> the lower layer and provide ease of use for developers and users. This
>> would also benefit desktop Linux users who run PA too (everyone running
>> Gnome or KDE) and make the option of running JACK and PA at the same
>> time
>> a more reliable and efficient process.
>>
>
> There are no real issues with running PA on top of JACK when you combine
> the right version of PA with the right version of JACK. I'm not sure what
> problems you think exist - it would be interesting to hear about them. As
> far as I understand it both PA and JACK developers consider this a done
> deal in each technology, but one subject to distribution/platform
> capabilities to make it work or make it fail.
>
>
>>
>> The discussion has so far been around what needs to be done to make it
>> viable for JACK to be included as an officially supported app
>>
>
> Work apparently. From your opening message in this thread:
>
>   "The optimisations that would enable PA to work better with JACK are not
> trivial and will require some pretty heavy thought and a fair sized chunk
> of time commitment to make them happen."
>
> It sounds as if you want someone to do some work on this for no other
> reason that mobile platforms are cool, or something. Oh, except it doesn't
> actually the MAJOR linux-y mobile platform at all. There is already a
> massive shortage of resources to put towards JACK development. Maybe some
> one new will take up your call to arms, and if so, great. Really.
>
> But looking at the "existing" JACK development community as if this is
> something that anyone here is obligated (or frankly, even likely) to be
> interested in ... well, to me it is just silly.
>
> You talk below about "various consortiums of large multinational companies
> that would be supporting this step" - why don't they create a paid
> position
> to do whatever it is that needs doing, away from the noise of this list.
> It
> would be more efficient, and more likely to get someone to actually do it.
>
> The thing is Patrick, you are talking platforms with tiny market share.
> JACK is a community of people focused mostly on music creation and
> pro-audio. The market you are talking about (in terms of devices) is close
> to meaningless. Linux isn't a big enough market already for this stuff
> (barely) and now you want to sell us on various secondary linux-y mobile
> platforms as the new big thing? Sorry, I'm not buying it. I'm not trying
> to
> be antagonistic, but you write about this stuff as if it ought to be
> bloody
> obvious to everyone how awesomely cool this is. All I see are more
> variants
> on meego or whatever that end up with essentially no market and certainly
> none of the excitement that exists within the iOS music creation/pro-audio
> community.
>
>
>> I disagree that in order to run JACK on the OTHER mobile OS's that
>> already
>> run PA by default we need to create and pay for a whole new audio system
>> that replaces both PA and JACK. That is the path that Google have taken
>>
>
> Actually, the seem to have gone further than that, unfortunately.
>
>
>> with both Android and ChromeOS and they have not succeeded with several
>> full time and well funded developers working on the task.
>>
>
> They don't understand the issues, at least it appears to be that way.
>
>
>>
>> However if the consensus from JACK developers is that they want to build
>> a
>> completely new audio system rather than spend any time on optimising the
>>
>
> Not a new audio system. All Linux based systems have access to a perfectly
> viable audio subsystem. Its name is ALSA. Mobile platforms just want to
> avoid exposing it (or even using it). Why? And why should we have to cater
> to that?
>
>
>> combination of PA and JACK then I will report that back to the
>> interested
>> and supportive parties who have opened the door to us and let them know
>> that unfortunately, due to political and personal objections it is very
>> unlikely JACK will ever be able to be officially supported to run on any
>> OTHER mobile platform than iOS.
>>
>
> It cannot run on iOS anymore. You missed the news, apparently.
>
>
>> Or is the hatred of PA so ingrained  that all JACK developers are
>> blinded
>> by their desire to see it vanquished from the earthly realm? The ongoing
>> quest for spiritual freedom of not having to run PA on any desktop
>> system
>> ever? The righteous quest of holy armagiddeon. The existential threat of
>> PA dominion. <rant>....</rant>
>>
>
> This is just stupid. I think PA is a pretty awesome system. In some ways,
> I
> may even be responsible for its adoption, having urged this very early at
> Ye Olde Linux Desktop Architects meetings back in the early 2000s. I have
> nothing but respect for PA. My respect is tempered with an understanding
> of
> its deliberate design limitations, which happen to be roughly opposite
> those of JACK.
>
> Instead of asking for volunteer help, why don't you try to actually hire
> someone?
>

As i said in the opening message there is a possibility that the Linux
Foundation could be interested in funding the development process if we
can come up with a viable plan. I have already committed my own funds to
funding myself to do the work required to get things to this point in the
process. Trust me when I say my wife is not happy about my decision to use
my time in this way. Its not just a case of sitting down and writing some
code. It also requires communicating with multiple "stakeholders" in order
to keep things progressing forward and gain insight into the issues that
need to be resolved. At any rate I am probably not the best person to
write the code that is required.

However while other communities have been very supportive, the results
round here so far have been people ignoring the actual topic of
discussion, focusing on Android development or denial of the viability of
optimising PA so that PA and JACK will work together more reliably on the
OTHER Linux mobile platforms that run PA by default and in the process
better on the desktop too.

In regards to the combination of PA and JACK being a done deal,
unfortunately there are several issues that have been exposed through
recent tests that have been run on the current stable versions of JACK and
PA. The issues are all inside of PA but they can be fixed if there is a
consensus or interest from JACK developers for such an effort. The fixes
are unlikely to come from PA developers as it is not exactly in their
priority list (or personal interests) to put their time/effort into
optimisations specifically to enable PA to run more efficiently on top of
JACK. Apart from that, doesn't it strike you as fairly unprofessional if
we require users to run specific combinations of JACK + PA in order to get
a completely reliable and stable desktop audio system?

- There is a bug in jack source/sink which allows the buffer to grow
continuously
- There is an issue with realtime capabilities in jack source/sink that
needs to be debugged
- It would be useful to make several optimisations to PA to enable even
faster realtime processes like optionally bypassing the Stream Buffer
while JACK is running but that will require some major design changes to
the PA graph and also some threading work to split out control and audio.

I know there are lots of people who advocate disabling PA in order to use
JACK but what about all the people who don't or cannot do that? Surely
they are the majority when it comes to desktop users and absolutely the
majority when it comes to OTHER mobile platforms that use PA as the
default audio management system.

So unless we want to go with option 2 which is to add and maintain support
for Bluez, Murphy and PA API in JACK are there people willing to work on
the PA optimisations that have been identified and if so what would be the
motivator to make it viable? Should we be looking to the Linux Foundation
or other well funded organisations to support us in this process? Can we
do it ourselves with the motivation that we can sell our products on these
OTHER Mobile platforms that run PA as the default audio management system?
Are there people who want to do the work just because they feel the urge
to contribute to the greater good and make open source desktop audio as
good as it can possibly be like in the good ol days before Linux became
the most popular operating system on the planet and the GFC began when
people had plenty of time on their hands to do awesome shizzle just
because they felt like it with no financial motivator (or was that a dream
reality that never really existed anyway)?




--
Patrick Shirkey
Boost Hardware Ltd

1381979990.8763_0.ltw:2, <58052.86.127.135.155.1381979981.squirrel at boosthardware dot com>