Re: How to determine database market share?

# Mark A wrote:
# Regarding the "relational model, I am talking about the definition
developed
# by Codd who invented the relational model. Hints and syntax to
improve
# concurrency did not even exist back when Codd developed the
relational
# model, so it is ridiculous to say that the relational model does not
say
# anything specifically about them. I absolutely guarantee that Codd
would
# puke all over hints and many of the other Oracle proprietary
enhancements
# because he considers them non-relational (to the extent they are
used).
#
# Even in the mid 1980's when Codd ranked databases to the extent they
were
# "relational," Oracle did not rank very high (way before they came up
with
# most of their current proprietary enhancements). My guess is that you
don't
# know who Codd is and never read any of his writings (he has been dead
for a
# number of years now).

Sorry Mark but I do know who Codd was.

The relational model is a "model" and it is more theoretical and abstract that what you are talking about.

I would agree that Codd probably would not like many of the things vendors have implemented.

However that is not because they conflict with the model. It might be that they conflict with what you are believing Codd would infer that the model implies certain things about how a vendor could ( attempt to ) implement the model.

It may be that you are not understanding the differences between a theoretical and abstract model and vendor's implementations.

The SQL standard ( various versions beginning way back ) don't match well against the relational model.

I go back to my earlier point that it's not a bad idea to talk about these features.

But to say that they conflict in the different manners you noted with the relational model when the relational model has no such concepts is not true. Received on Sat Nov 05 2005 - 21:43:41 CET