It's powerful, very easy to customize, extremely fast, and handles huge XML documents quite well :) And besides, it isn't simply outputting "Hello World." It's recursing through various structures/nodes to duplicate an entire document, allowing fine-tuned controls with the simple touch of XPath. Hurrah!
I'd be curious to see how it benchmarks against some of Perl's XML modules. Perhaps it's overkill for something as simple as comment removal, perhaps not.