On Sep 26, 2005, at 3:43 PM, Jeremy Dunck wrote:

> Tangentially, ajaxian connection profiles mean a different kind of
> scalability is needed in web servers.  This has a reasonable
> discussion:
> http://www.mortbay.com/MB/log/gregw/?
> permalink=Jetty6Continuations.html

Ah, interesting, thanks for the heads-up.


> Greenspun's Panda Ch. 10 and 11 outline the distinctions well.  Django
> is a CMS first.  Rails is an app platform first.  ASP.Net is a kitchen
> sink with squeaky knobs.  ;-)

Thanks too for the reminder, I hadn't looked at that guide in several
years.  Your assessments sound like what I've heard elsewhere.


>> Or are you
>> seeing the kind of development-time speedups these things (and the
>> fawning
>> hype surrounding them) promise?
>
> You mean like "10x more productive"?  Well, just an ORM will go a long
> way to saving time, and model-from-DB or DB-from-model are both nice
> approaches.  And the culture of unit testing is also a force
> multiplier.  But see here for a strong rebuttal:
> http://fishbowl.pastiche.org/2005/03/11/catching_a_silver_bullet

Indeed, the essential/accidential split is good to remember.  As is
being pointed out in #code4lib, too, the implementation vs. iterative
improvement split is also important, and you don't know until you've
tried it.

I suppose that with those both in mind, I'm even happier with the
quixote+PTL combo I've been using, as I can usually add new features
today (18-24 months since start of development) in roughly the same
magnitude of time it took to add initial features.  That says a lot,
even if changing/extending things takes a little longer.

I'm still going to try TurboGears on my next "little" webapp though. :)

   -Dan