[Twisted-Python] web vs web2 clarification

Mikhail Terekhov termim at gmail.com
Thu Dec 10 15:29:09 MST 2009


On Thu, Dec 10, 2009 at 2:41 PM, Phil Christensen <phil at bubblehouse.org> wrote:
>
> I think that's debatable. Mozilla was terrible for a long time, there just wasn't much alternative. You could make the (arguable) point that Mozilla's rewrite happened at the cost of disengaging from the web community, leading people to jump ship to IE.
>
I agree that the process wasn't smooth, easy and done in a best
possible way. But the point was that the net
result is positive. Would "Mozilla Platform" be possible _now_ without
such a rewrite?

>
> I don't think Apache is a particularly good example of why ground-up rewrites are more reasonable in open source projects. It took **years** to get a critical mass on Apache 2, and there are still many plugins that don't support anything but prefork mode.
>
I meant here the transition from 0.7.x to 0.8.8 (Shambhala). They
designed it completely from scratch in
_parallel_ with developing the 0.7.x series. And this was a foundation
of the following 1.x success (modular
design etc.)

>
>> And what about CVS that couldn't/refused rewrite from
>> scratch and peacefully evolves adding features little by little? Can it
>> catch up with svn/git/bzr/hg? I doubt so. CVS's example in fact says that
>> if you refuse to do it then someone else will do and will replace you.
>
> Well, again, debatable. OpenBSD still uses CVS, and has started a significant push to patch and upgrade the existing CVS source.
>
With all due respect to OpenBSD developers, they are just a tiny part
of the all VCS users and their
significant push in their relatively small community will not
magically save CVS I'm afraid. The point here is
that once a dominant VCS in the world, CVS now looses its position
pretty fast and all these incremental
improvements (commitid etc.) can not save it. It is too late now. The
part of the lesson here is that the
code is/was in such a shape that the implementation of urgently needed
features (like commit atomicity)
took so long that others (svn/git/bzr/hg etc.) implemented everything
from scratch and even more.
See for example of VCS usage trends http://tinyurl.com/ykmurn7 and
http://tinyurl.com/qj7c3

>> May be it is open source that makes it different, may be people think
>> more about fun and beauty than about money in this case?
>> It is not so simple and Joel's hypothesis is far from being 100% true IMHO.
>> In the end a lot of people know/use Borland's, Netscape's and Microsoft's
>> software but how many know/use Joel's? ;)
>
> I think it's far simpler than that. There's an old expression that sums it up: A bird in the hand is worth two in the bush.
>
> Taking limited resources from a nearly-complete and reasonably popular project to add them to a brand-new and **backwards-incompatible** project is rarely ever a wise project management decision.
>
That is all true but it is very close to Joel's reasoning, kind of a
manager's point of view.
It is too business/money oriented and doesn't exhaust all the reasons
why people write
software in open source world in particular. And what is more
important it doesn't explain
why they still rewrite it from scratch sometimes and succeed? ;)

But this became completely unrelated to this list, sorry for dragging
discussion so far.

-- 
Mikhail Terekhov




More information about the Twisted-Python mailing list