Mark is very energetic guy and nicely talks about editable XPathDocument as a preffered XML store in .NET2 (has been cut, stick with DOM for several more years) and of course XQuery implementation in .NET2 (cut). So basically all new XML features that still not cut are about performance, some API rearranging and XSLT 1.0. Performace was an old pain and had to be fixed anyway. XSLT 1.0 stuff is good - finally XSLT debugger (which actually was hidden as internal stuff within System.Xml.dll since .NET 1.0 - just run Reflector to see it), better XML and XSLT editor in Visual Studio .NET 2005, and brand new XSLT 1.0 processor, which compiles XSLT down to MSIL - this one is just perfect and I'm looking forward to give it a whirl. So as a matter of curiosity .NET 2.0 is going to include two different XSLT 1.0 processors (one obsoleted though).
Well, I'm not saying that's nothing. But let me be harsh - it's all catch up stuff. It would be excellent to have in .NET 1.1, but for year 2005 it's quite disappointing. May be that's because all that hype about XQuery and how it's better than XSLT. There were really hot debates on Microsoft decision to implement XQuery, but not XSLT 2.0 and now guess what - neither XQuery nor XSLT2.0 and developers running out.
I can imagine how much resources have been spent on XQuery! That's not a small one. Apparently managed XQuery turned out to be just another black hole project. Who knew, right?
Oh well, there is Saxon.NET and may be will be XQP. But I fell a little bitter taste when I realize I'll still be helping people struggling with XSLT 1.0 limitations at least next 3 years, while XSLT 2.0 is shining in a better world. And my fellow XML MVPs are sharing my feelings. Still - what a sad irony - to listen to Mark, who has left XML team talking about XQuery impementation, which was cut...