The benchmarks were from what was designed to be a fair test. It might not have worked out that way, but that was the intention of the creator nonetheless. I stuck a massive disclaimer on my link to those benchmarks and let people at it (since NO benchmarks are EVER undisputed - biggest lesson I've learned). The post is at http://www.rubyinside.com/ruby-xml-performance-benchmarks-16... if you want to see how it was portrayed.

If everyone had to bother validating all the third party bits and pieces that get referenced on blogs, no-one would blog. Blogging is a "hey, check this out, I ain't saying it's true, but you might find it interesting" type of affair - it's not the New York Times (which is why regular journalism is foundering; it's expensive to fact check everything and, heck, it's a Ruby blog, not a trusted source of journalism).