On Mon, Aug 11, 2008 at 7:02 PM, Tim Spalding <[log in to unmask]> wrote:
> I see errors in both directions. There are some CRUD pieces to
> LibraryThing that took *way* too much time to do because we did them
> all by hand. Casey's introduction of a Django framework there was very
> smart. At the same time, I think Twitter got stuck in RoR. When a web
> application is so extremely simple--it does basically one thing!--you
> might as well write that as "close to the machine" as possible.
To be fair, Twitter's problem is less the fault of RoR and more the
result of poor database design/planning. I'm no Rails apologist
(after all, as we all know, it's a ghetto - it, like every other
stack, has its pluses and minuses), but the beating and FUD Ruby on
Rails has received thanks to Twitter's Fail Whale is largely
undeserved. Would it be more stable if it was some
Spring/Hibernate/JSF application? Who knows, but it's doubtful. If
the problem really *was* all with ActiveRecord and not with their
underlying database architecture, their problems should be easily
solved (with the word "easy" meaning "easier than the drubbing they've
received from their critics") by exposing their DB with another stack
and porting their models over to ActiveResource.
So, really, these are two completely separate issues.
**Full disclosure -- I am an unabashedly Ruby-centric developer,
although generally avoid Rails.