Which raises an interesting question. Apparently, browsers will take an XML file and a link to an XSL transformation and generate HTML (don't know know if it validates, though), which raises the question of whether to bother transforming it server-side. Then again, most search engines suck at searching arbitrary XML, particularly in the absence of a schema, which might be a publisher consideration.
I'm going to spend some "quality time" reading up on XML, XSL and related technologies. The .NET Framework has a robust set of XML processing functionality which I'll continue to play with.