On Mon, Jun 15, 2009 at 10:37:04PM +0200, Stefan Keller wrote:
> @David: You wrote in the links cited "The "flexibility" stems from
> fear of making a design decision.". That's an important note.
> Nevertheless, there are use cases where you *can not* know in
> advance what the name is of the attribute!
Those cases are extremely rare, and they don't fit with an RDBMS.
> To me that's not fear but adaptiveness, modesty and knowing when to
> break the rules! An apparent successful example of this EAV design
> is OpenStreetMap
> (http://wiki.openstreetmap.org/wiki/Database_schema).
Funny you should mention that. The OpenStreetMap people are hitting
exactly the issues I named, and they're ruing the day they decided on
that "schema." I don't recommend that anybody re-make their mistake.
Cheers,
David.
--
David Fetter <david@fetter.org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david.fetter@gmail.com
Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate