As much as I appreciated Jim Wooley's response to a post on the xmlteam blog ...
XLINQ is so simple that it doesn't need much discussion. (The same can't be said for the XML Dom which is another reason why I think XLINQ rocks)
... I hope people don't just trust us to do the right thing with XLinq. We really want people to try it out, yell at us for things that aren't clear or don't work right, and let us know what DOM or XQuery or XSLT or JDOM/XOM does better.
In the six months or so that I've been the program manager for XLinq, I've seen a lot of really beautiful designs go CRUNCH! against the nasty rocks of Reality. Lots of bad things happen in the XML specs and in XML as practiced that make some of our ideas look a bit pathetic in hindsight. People who hope to use XLinq someday can't assume that our little team will catch and fix all the gotchas in the all-too-few number of weeks left in the Orcas development cycle.
PLEASE try XLinq (and the other LINQ components, especially in combination) to prototype some realistic projects with real data, and let us know your experience. We don't even hope that it will do all things for all XML people, but we do hope that it is easy and efficient to use for most mainstream use cases that involve XML creation and processing in the .NET environment. Preliminary feedback indicates that we're going in the right direction, but I'm sure there are a lot of devilish details you can tell us about that need attention.
[fixed spelling of Jim Wooley's name, sorry about that!]