Does anybody here have experience reading the SGML/XML files representing the content of EEBO? 

I’ve gotten my hands on approximately 24 GB of SGML/XML files representing the content of EEBO (Early English Books Online). This data does not include page images. Instead it includes metadata of various ilks as well as the transcribed full text. I desire to reverse engineer the SGML/XML in order to: 1) provide an alternative search/browse interface to the collection, and 2) support various types of text mining services. 

While I am making progress against the data, it would be nice to learn of other people’s experience so I do not not re-invent the wheel (too many times). ‘Got ideas?

—
Eric Lease Morgan
University Of Notre Dame