[zeromq-dev] HA: I lost a bet
Michel Pelletier
pelletier.michel at gmail.com
Wed Apr 25 19:48:08 CEST 2012
Ok here's a crazy suggestion. I think what would work really well is
a sort of advanced 0mq boot camp. There are obviously highly skilled
engineers in our community who could easily take up some of
understanding that until now was contained in only one or two people's
heads.
I would love to get a high level overview of the various internal
parts of the library, the moving pieces, the internal interface
points, the memory allocators, and the critical structures. I've done
code analysis but without context it gets hard to see the big picture.
I'd be willing to fund my own travel and lodging to a central place
where we can get together and do this.
I should note that unlike the conference event or meetups, I'm
speaking specifically of focusing on core code details. A proficiency
in engineering and C++ would be required, as well as up-to-date
knowledge of the current 0mq 3 API including all socket types and
functions. I don't want to exclude anyone, but time would be wasted
answering questions like "what is pub/sub?" We'll save that for next
year's conference.
Any takers?
-Michel
On Wed, Apr 25, 2012 at 9:43 AM, Pieter Hintjens <ph at imatix.com> wrote:
> On Wed, Apr 25, 2012 at 6:20 PM, Sergey Hripchenko
> <shripchenko at intermedia.net> wrote:
>
>> Sorry to touch these sensitive things, but just wanted to comment that still you will need someone to be an architect and to know the deepest ZMQ implementation details.
>> The most valuable help for my commits was from Martin off the list.
>>
>> I found some issues with ZMQ and still don't know whom to ask/talk about it.
>
> Sergey, the only way these things can work is that people like
> yourself with the talent learn from those like Martin with the
> knowledge. If it's too hard to learn, or there are barriers, then that
> knowledge remains in one person's mind (which has been our story up to
> now). Which is not safe or sustainable.
>
> Spreading the core knowledge is one goal I think we can agree on.
> Having a wider set of contributors also means we focus on the right
> problems and do them better (like your unbind/disconnect work).
>
> -Pieter
> _______________________________________________
> zeromq-dev mailing list
> zeromq-dev at lists.zeromq.org
> http://lists.zeromq.org/mailman/listinfo/zeromq-dev
More information about the zeromq-dev
mailing list