Re: [Jack-Devel] The Situation(s) With JACK
On 15/12/11 00:02, Stéphane Letz wrote:
>>> And one way to improve the situation is to make a clear choice
>>> or the reference code base.
>>
>> Paul replied:
>> i don't think that *anyone* believes that the C code base of JACK1
>> is an appropriate thing to keep developing. certainly I do not.
So it seems that JACK1 is now "officially" dead. That "solves" a
lot of potential problems and arguments.
>> the questions is: (given that) we (active JACK developers all)
>> want/need a C++ implementation. is the JACK2 codebase the right
>> starting point for that, or do we gain something by doing a new
>> implementation that draws from JACK2 but leverages boost (among other
>> things) and uses fully idiomatic C++ unless doing so compromises
>> performance notably.
>
> Stéphane continued:
> OK, assuming the answer to "do we gain something by doing a new
> implementation" is yes, the next question is : who is going todo
> that? who is going to do multi-OS support? who is going to port
> current specific code (backend, bridges?) to this now code base?
> who is going to debug/test it?
It seems the quasi-consensus is that a total rewrite would solve many
current and future problems. Unlike for both JACK1 and JACK2 there is
now a decade of JACK usage and history to leverage for a redesign but
the man hours required is huge so plan on the best possible intermediate
compromise with JACK2 while laying the foundation for JACK3.
As a long term interested non-developer the obvious path forward to
me is that whatever functionality that JACK1 has that JACK2 does not
needs to be ported across, if possible, and then use JACK2 as *the*
canonical reference codebase for the next 12 to 24 months that it
will take for the planning and initial deployment of JACK3.
1323877867.27579_0.ltw:2,a <4EE8C5A3.1070508 at renta dot net>