Key takeaway here: Ruby on Rails is the new, hot fad, but apparently doesn't scale in the way web languages should. Dropping ActiveRecord, for example, cuts out the main piece of RoR coolness I still liked.
Disagree. It seems more like they've optimized one part of their infrastructure, so now a new bottleneck has popped up. Par for the course when you're optimizing anything...
Oh, and this interview was discussed by DHH on his blog:
The comments were interesting, especially one from Alex Payne (the Twitter dev himself). Here is an excerpt. Visit the link above to read the full comment (which is kind of long to post fully inline here).
"...We'd be running into scaling issues on Django or any other platform. We're happiest scaling on Rails."
Seems like RoR is great for getting something out very quickly but if the site takes off then you need to be prepared to take out the very features that helped get you there.
So, it comes down to, if you're shooting for something very high traffic (and who isn't) use a faster language. If not, use RoR.
Django gives you the benefits of speed and an easy to use framework. Rails just seems to have more hype around it.
The people who constantly say that the poor performance of Ruby and Rails isn't an issue are the same people who don't have large spikes in traffic, since their apps are subscription-based. For most of the typical Web 2.0 stuff getting built these days, huge spikes in traffic are exactly what you want, but using Rails is setting you up to have scaling issues once you get the press you're hoping for. You're likely to run into those same issues regardless, but why make things worse? Assuming Rails is better than Django, it's not so much better that it'll make up for the poor performance. The improvements in ease of development aren't unique to Rails, they're a byproduct of the philosophy that Django, Rails, and similar frameworks all share.
I'm far more likely to trust the opinion of the Twitter folks than that of people who have something to gain from my choice of framework.
I agree with you. I've been debating whether to focus on using Django or Rails - and I think I'm going with Django. If Rails has known speed limitations, why work with it if there's another option?
