From: RB
Date: 28 Jul 2005
A few questions, thank you for bringing some insight on the
following points:
When you state that "OODBMS" do not have any
theoretical foundation, do you consider they have better physical
implementation than SQL DBMS ?
The reason I ask the following is because I have set up my
own simple tests with equal functional and applicational stress tests on a SQL
DBMS (SQL Server 2000) and Intersystems Caché 5.0. The results allow me to conclude this: Caché is a better physical implementation for retrieving data simultaneously while making complex
computations on the fly. I believe this is the source of lot confusion I have
read. As "ODBMS" seem to be better physical implementations than SQL DBMS, they induce people to think they would have a better model behind which they don't. They seem to do better in the "HOW" but not in the "WHY".
3 years ago, I had met a 58 yr old Australian DBA who
encouraged me to educate myself on SQL and RM fundamentals. He told me that SQL was meant to be better than it is and was somehow "sacrificed" for budgetary reasons (he
stated that he worked on its elaboration which left him with deep sense of
frustration). Later on, he decided to
develop his own DB engine (named ATLAS) and his own version of SQL he considers
a better implementation of RM. I really would like to have your opinion about
this paper who's supposed to be an explanation about limitations of SQL and how
he gets around them. It is a techno
business paper to help promote his product but it's probably the most
informative one I have found about a software product. Here's the link:
http://armadillo.fr/english/whitepapers/WHITEPAPER_2004.PDF
From: Fabian Pascal
You are asking the wrong question using the wrong evidence. I
cannot do it justice via email. All I can say is continue to educate yourself
and you'll see where you got it wrong. Pay particular attention to how
"equal functional and applicational stress" is defined, and
include in your criteria human, not just machine performance; integrity,
not just manipulation; maintenance, not just usage; soundness,
not just speed of response.
That is only one of the confusions prevalent in the industry,
and it's there due to lack of proper education on fundamentals.
One look at the first couple of paragraphs of the paper you
mention tells me he has no clue what the RM is. It is true that SQL went wrong,
but his perspective cannot be but nonsense.
Posted 9/30/05
