Can anybody give me a good summary of the difference between an object database, an object-relational database, and PG? So far, I've figured that object databases seem to be efficient persistence systems for OO languages. Trouble is, if you're using some OO languages, and some non-OO languages, you have a problem. Also, it appears that with most OO databases, there is very little power within the server; no objects like stored procedures, or rules, or views. If you want to do anything, you have to instantiate your data, which is great in theory, but hell of a slow in implementation.
The commercial object-relationals that I've had a look at seem to add on the object bits almost as an afterthought (I guess that's pretty much what they are), and don't seem to have core object functions. Or am I just missing them?
PG seems to fall into a category all on it's own, as it seem to have better object support than all the other object-relational vendors, yet better support for processing and control functions than the commercial object DB vendors.
What am I missing?
MikeA