Re: Sensible and NonsenSQL Aspects of the NoSQL Hoopla

From: <karl.scheurer_at_o2online.de>
Date: Sat, 31 Aug 2013 08:22:41 -0700 (PDT)
Message-ID: <c8d8012c-bc93-4e7d-9627-b378e2534e71_at_googlegroups.com>


Am Donnerstag, 29. August 2013 00:54:31 UTC+2 schrieb com..._at_hotmail.com:
> On Monday, August 26, 2013 3:20:44 AM UTC-7, karl.s..._at_o2online.de wrote:
>
> > What is the benefit of "logical model" that cannot be mapped lossless to
>
> > a physical storage? Sorry, but relations are abstractions of "record
>
> > oriented" storages.
>
>
>
> You have a misconception of the relational model. A relational schema is >intended to express what you call a "conceptual model". A relation variable or >query has an associated predicate and its tuples are those that make true >propositions/statements.

Sorry! I did not phrase my critics in a clear and comprehensible way. Based on a background in algorithms I see no way to implement the relational model in a complete and lossless way. Every "real" relational system is like a car with 5 gears using only 3. Implementing relations as a kind of objects is no difficult task and been done several times. Accepting the object nature of relations can be the difference between success and failure. My problem is not the usability of the relational model, but the irrational hold on as a complete answer for all database problems. Codd's model emerged out of the technology of the seventies and needs urgently a revision.

m.f.G
Karl Scheurer Received on Sat Aug 31 2013 - 17:22:41 CEST

Original text of this message