This is quite an interesting idea — benchmarking feed parsing libraries in different languages. I'll give it some thought.