I’ve spent a couple of weeks wrestling with Nokogiri to parse a tonne of DATEX II data in to some usable format. Previously I had a mash of libxml and REXML, and the code was either ‘fast’ or ‘pretty’, but not both.
Nokogiri is good – it’s very good, in my opinion. The only trouble is, documentation and examples are a little thin on the ground, which slows everybody down. Here’s the dilemma – do I spend time writing poor documentation based on my limited understanding of part of Nokogiri, or leave it to somebody else?