This is greatly appreciated (though I'm not going to look at it tonight, seeing as it's about bedtime for this early riser). The database agnosticism you're suggesting is of course of great help to the OpenACS project, so I look forward to reading the spec and providing feedback.
