My understanding is that mainstream databases are kinda like mainstream economics -- dresses itself with the trappings of math, we're supposed to take its practitioners seriously, but it turns out to be clunky and unnatural at actually doing its job. ;) <http://home.pipeline.com/~hbaker1/letters/CACM-RelationalDatabases.html>
When I last looked at the dominant model, it didn't seem that many vocal supporters (like Chris Date) fully understood the alternatives.
(Personally, what has always weirded me out about common databases is they look nothing like what people typically do with data. People stuff their models into this unnatural form, for little compensation.)
This weirdness is pervasive within the tech world. It's amusing to look at job ads and see how near-unemployable I am for many positions. The clumsiness of most mainstream tools is too great. Many others are in the same boat. <http://video.google.com/videoplay?docid=-2950949730059754521>
Tayssir