Good point. The last thing we want is another Python vs RoR vs whatever war. I know that I helped test performance and ran my own benchmarks when Massimo was working performance improvements with regards to session files back in March, since sessions can slow down response times. My own testing found that my app, which is becoming quite large now, had an average response time of 30 ms (without database interaction, of course).
I don't care what language or framework you use. If you can get a complex application (with db turned off and no performance tweaks) to give average response times of less than 50 ms, then it doesn't matter how the framework is rendering the files behind the scenes. I've never used RoR so I don't know if it truly is faster or slower, but even if it manages to shave a couple of milliseconds off the average response time compared to web2py by caching everything in memory, does it really matter? Your average response time is still well below 50 ms which is FAST! You will never see a difference between 30 ms and 28 ms. I doubt even Google or Facebook would care about a 2 ms difference. At that point, just use whatever you are most comfortable with.
