1 Feb
2016
1 Feb
'16
2:53 a.m.
On Sun, January 31, 2016 3:13 am, Paul Davis wrote:
Sometime in the next two weeks, I will find the time
to deal with a
variety
of pull requests for JACK 1, update some articles on jackaudio.org
(notably
FAQ stuff), and do a new release of JACK 1.
This will be my last work on JACK. The time has come for me to step down
from my role as "benign dictator (and jack1 maintainer)". There several
reasons for this:
* most linux distributions use JACK2 as their default, so JACK1's
relevance has diminished. I
still believe JACK1 to be a superior choice from some technical
perspectives, but there is
no doubt that JACK2's integration with dbus and thus its
interoperability with PulseAudio
has made this the safe and simpler choice for Linux.
* I really don't have the time to even think about things related to JACK
these days. It does
any future development a disservice to have me as the bottleneck, which
I effectively am
at the moment.
* Because 110% of my time is spent on Ardour, the fact that Ardour now
has
non-JACK
audio/MIDI I/O options has diminished the significance of JACK for my
own work.
* as the years have gone on, although I am still delighted by the
technical quality and
the conception of JACK, I no longer think that it is a particularly
good
idea for most users. There
are times when it is useful
I will continue to pay for the hosting of jackaudio.org (even though JACK2
continues to be distributed, managed and communicated about via other
channels), although if someone wanted to migrate this to some other more
communitarian platform, we could look into that.
I would be happy if someone volunteered to step up as maintainer of JACK1.
It would obviously be even better if someone was willing to take the big
leap to JACK3, a version that combines all the best parts of JACK1 and
JACK2, but I think it is more realistic to accept at this point that this
is not going to happen.
If nobody does step up, then there is a good chance that JACK1 will become
officially unmaintained. This isn't of much consequence, because once the
latest pull requests are merged, there won't be any known bugs in the
code,
and also because not many people use it anymore. This also means, of
course, that "maintainer" is not much of a task, should someone feel
hesitant about taking it on.
The end of an era!
A big congratulations to Paul and the Ardour team for getting to the point
where Paul is too busy to work on JACK anymore. That is a great
accomplishment and is due in large part to the original development of
JACK to create a solid foundation for Ardour.
JACK1 is still a very useful part of our toolkit. We routinely use it for
distributed audio processing on a range of machines with many different
versions of Linux. In that environment it seems to be more stable than
JACK2.
It seems that the differences between JACK1 and JACK2 relate to the
intended target use case. JACK2 does a decent job of trying to enable many
consumer friendly tasks and take out some of the pain of doing inter
application audio production. JACK1 handles some heavy lifting scenarios
more effectively than JACK2.
Maybe that is a good direction to continue for the different codebases?
JACK1 - Heavy lift mechanism for high performance Linux platforms
JACK2 - Flexible user friendly solution with cross platform support
I don't see any problem with continuing the split development process in
that regard. Maybe we don't need a single tool to do everything anyway?
I hope that Paul will continue to provide his insight and expertise as
part of the JACK development community.
If he is interested, I nominate Harry Van Haaren as the new official
maintainer for JACK1 codebase. He has proven his skills and commitment and
is a fair and impartial communicator.
Anyone else want to put their hat in?
--
Patrick Shirkey
Boost Hardware Ltd