[noir_lord]

The inevitable Schema vs Code mismatch that creeps in.

I've yet to see a belts and braces solution that works when introducing it on an existing database (rather than from the start with something like Flyaway).

I mostly like SQL but I kind of wish someone would do a clean slate design of a relational data store and unfuck the last 40 odd years of cruft.

I work in an environment where our data is eminently relational and generally mapping to a relational db type structure is quite simple but actually dealing with the minutiae of the various RDBMS's is just painful.

[tluyben2]

I was wondering about this a few days ago when I was, again, stuffing a square peg in a round hole by doing work you would hope a db or orm should be able to handle.

Is there any research going on to solve this in a non OO environment? I worked with (expensive) OO databaes in the past and that was crap, nosql is not helping much either, although the idea (I say the idea as I hate the implementation) to have the language married with the datastore like Mongo is something. I have seen non OO environments where the company just extended sql to be a full generic programming language which also works but it all feels... Not quite there. Q/k on kdb+ I find pleasant as I find Mnesia with Erlang. That feels far more natural than the other choices, but I feel it can be better with research and more FP oriented.