2011/3/9 Jeremiah Foster <[email protected]>:
> On Mon, Mar 7, 2011 at 5:09 PM, Arjan van de Ven <[email protected]> 
> wrote:
>> Hi,
>>
>> given the events of the last few weeks, the MeeGo architects have, and still
>> are, revisiting various parts of the MeeGo architecture.
>
> Aside from this thread, where does this discussion happen? How do others 
> engage?
>
>> Having said that, three items are currently clear enough to make a final
>> decision on:
>>
>> 1) MSSF / MeeGo security framework
>> 2) Buteo Sync
>> 3) PIM storage (currently stored in the tracker database)
>
> Reading between the lines, it appears that the MeeGo architects are
> re-evaluating Nokia contributions. This is well and good, but the
> question that you do not address Arjan is who is going to be
> responsible for assuring that this re-evaluation is done with careful
> consideration of both MeeGo's stated target architectures? If you rip
> out libraries and software that is optimized or designed for the ARM
> platform for example that may lead to the perception that MeeGo is
> sliding back in Moblin.

Pitching in with my opinion. The areas mentioned by Arjan does not
affect anything specific to a certain platform, but to middleware. The
good thing about MeeGo is that in practice, MeeGo on ARM and X86 acts
the same. If we design software and put it in MeeGo especially for ARM
(or X86), we're doing it wrong, unless it is part of a higher
abstraction that reduces it to be hardware adaptation that has a
counterpart on 'opposite' side.

> Careful assessment of the current architecture requirements needs to
> be done from a cross-platform perspective and done in the open. It
> would be good if MeeGo had more ARM experts at the architects level to
> ensure parity in at least discourse on design. Right now it is in
> danger of sinking into a back room monolog.

I personally think this could be cooked down to "Architecture needs to
be done in the open" - ARM experts, customers of the MeeGo platform,
X86 experts or whoever can then weigh in. Transparency and meritocracy
should make sure that proper decisions are made and discussed.

It is however important to remember that in the end, -someone will
have to make the call- - architecture can't be a democracy or we could
be discussing endlessly and would be going nowhere. Arjan has stated
what needs to be done in the short term - we need to get MeeGo 1.2 out
the door in a proper shape.

It is more important than ever now to discuss how to do proper
architecture process in the open. Decisions on architecture -is- a
power position but also one of big responsibility. It determines
direction of the actual code in the project. And that's why it's one
of the most important functions to have properly open in the project.
So at least contributors know how MeeGo is supposed to be put together
and that it functions properly.

What do you (architects or even want-to-be architects, please do weigh
in as well) think would be a proper, simple, understandable
architecture process that you'd be willing to participate in?

BR
Carsten Munk
_______________________________________________
MeeGo-dev mailing list
[email protected]
http://lists.meego.com/listinfo/meego-dev
http://wiki.meego.com/Mailing_list_guidelines

Reply via email to