Re: EAV (Re: Object-relational impedence)

From: Dmitry A. Kazakov <>
Date: Sat, 29 Mar 2008 14:58:32 +0100
Message-ID: <1j95sy6k5mjy2$.suyl1sbc5as0$>

On Sat, 29 Mar 2008 13:14:49 GMT, David Cressey wrote:

> "Dmitry A. Kazakov" <> wrote in message

>> It is interesting to see people reinventing the wheel (class). The set of
>> "all possible widgets having attributes T, U, V" is a class.

> The pattern of "generalization-specialization" predates both the relational
> data model and the object oriented paradigm. Neither one of them really
> derived that pattern from the other. There is undoubtedly a common
> intellectual ancestor, but you might have to go back centuries to find it.

Yes. That by the way implies that neither is a paradigm in that respect. "Class" is just an implementation of the pattern, i.e.

  1. how could I describe a set of similar types? (Similarity is defined in terms of behavior.)
  2. can I have make its closure another type? (This gives polymorphism)

The same steps would apply to relations. So generic relational programming (if that thing will ever happen) would deal with classes of relations represented by relations.

Dmitry A. Kazakov
Received on Sat Mar 29 2008 - 14:58:32 CET

Original text of this message