I was at Froscon last weekend and it really was an interesting presentation. But using the example of a simple “Hello World” output to measure the performance of the frameworks, was a classical type of “never believe a statistic you have not faked yourself”.

I’m sorry, but if some of those frameworks perform anywhere near the level in those slides on something as dead-simple as “Hello World”, that’s just awful.

The truth lies somewhere in the middle …

I have to say, I’m not all that suprised at the preformance of those frameworks. I don’t know them all, but the mayor ones I’ve looked at are all really bloated. (But I gues it is all in what you expect a framework to do.)