I think the concept of DBIx::AnyDBD is great and works well (all though problems do arise arounf ODBC / ADO/ OLEDB but that is not Matt's fault). As long as it does what DBIx::AnyDBD did that's fine by me.
What I like is the fact that it will be integrated into core DBI and promote the use of this subclassing mechanism as a basis for writing cross-platform code rather than each modules reinventing the wheel (c.f. Alzabo). -- Simon Oliver
