I used an XML DB in a project back in the early 2000s. The database was populated by running an in-database XQuery that pulled data across the webs in XML. When

I used an XML DB in a project back in the early 2000s. The database was populated by running an in-database XQuery that pulled data across the webs in XML. When people visited a web site, XML was queried back out, then XSLT translated it to (drum roll) XHTML. I even experimented with XForms to capture data from visitors, which had the benefit of directly using the DB schema to validate data. All in all, it was pretty nifty when you got all the pieces working together properly. Sort-of XRX without the R. As to reasons for their "death": DTDs and XSDs were a serious PITA. XML data types didn't map closely to standard relational DB types; this wasted space (which was costly at the time), and made it difficult to interact with a relational DB when that was necessary. The XQuery query planner was braindead. That was really the nail in the coffin. All in all, they're just document stores, and those are more popular than ever now with the whole NoSQL thing. As the author mentions, XML seems to have won out as a document format... I wonder what sort of DB we could put all those documents into. One that could support storing those documents without a bunch of translating to other formats, and had some method of running queries over them (preferably in a standardized way), and output the data in a similar format. Too bad XML databases are dead, one of those would have worked pretty well.