I'm trying to read through the documentation on Berkeley DB XML, and I think I could really use a developer's blog post or synopsis of when they had a problem that found the XML layer atop Berkeley DB was the exact prescription for.
Maybe I'm not getting it, but it seems like they're both in-process DBs, and ultimately you will parse your XML into objects or data, so why not start by storing your data parsed, rather than as XML?
Ultimately I want my data stored in some reasonable format.
If that data started as XML and I want to retrieve it/them using XQuery, without the XML layer, I have to write a lot of code to do the XQuery by myself, and perhaps even worse to know my XML well enough to be able to have a reasonable storage system for it.
Conversely, so long as the performance of the system allows, I can forget about that part of the back end, and just worry about my XML document and up (i.e. to the user) level and leave the rest as a black box. It gives me the B-DB storage goodness, but I get to use it from a document-centric perspective.
