Re: arch, svn, cvs
* Christoph Hellwig:
>> Greg Hudson contributes an interesting viewpoint:
>>
>> <http://web.mit.edu/ghudson/thoughts/bitkeeper.whynot>
>
> It's completely unfounded bullshit. Whether you prefer a pyramid
> or lots of commiters style organization is pretty much a personal
> or rather community organizational issue. Both have their up-
> and downsides.
Of course, but if I look at Linux release management (which is still
somewhat problematic, as countless vendor-specific patches can attest)
and cursory pre-integration peer review, I think he has a point.
Is there *any* large project using a pyramid model and produces
regular, high-quality releases which are used mostly unchanged by
distributors?
> All the actual arguments he presents are totally unfounded, abd
> besides that bitkeeper as the specific SCM he mentions supports
> multiple commiters style development just fine - for bkbits.net
> projects for example you can add multiple ssh keys and thus let
> different people push to the same repository.
The BK-specific criticism is the less interesting part, and Greg
himself mentions that BK can support centralized development. His
remarks on the pyramid model are the relevant part.
Reply to: