Re: Sensible and NonsenSQL Aspects of the NoSQL Hoopla

On 02/09/2013 1:07 AM, karl.scheurer_at_o2online.de wrote:
> Codd shifted the complexity from one area to another. When dealing with
> n entities is replaced with n*x relations is a considerable increase in
> complexity. One reference told me that a SAP R3 system contains more than
> 100000 tables (one hundred thousand!). What about complexity?

No. Codd shifted what was common in programs (programs of the 1960's, not the 1970's by the way) to the dbms. Obviously no one person nor any manageable group of people can decide how redundant SAP is, so SAP must be regarded as a failure as far as Codd's idea is concerned. If your statistic is accurage, SAP must be regarded as just as much a failure as the systems it seeks to replace.

James Lowden is right that the obvious programming leverage of the RM is in removing data manipulation from program text, eg. through declarative views. The DB world has generally failed at this too. The answer is obviously not more education of application programmers, not more data manipulation in app program text. Eg., most application programmers have been so cultured that they cannot imagine how to function without, for example, assignment. The general mistake is failing to recognize that there is more demand for programming talent than there is programming talent. That's true in every field but it's especially so in IT which is a much younger field than most.

If part of your meaning is that some things today would be much different had we had today's hardware in the 1960's, I agree with that.   For one thing, there was no commercial online database then that couldn't be recorded in today's RAM. Received on Mon Sep 02 2013 - 17:13:42 CEST