Re: A Question On Many-To-Many Linking Table(s)

From: --CELKO-- <71062.1056_at_compuserve.com>
Date: 25 Sep 2002 11:03:58 -0700
Message-ID: <c0d87ec0.0209251003.74abc342_at_posting.google.com>


>> Umm, I'm guessing you don't know much about programming. <<

I have been in the trade for 35 years so far, taught college and taught industrial courses.

>> A pointer is called a pointer or a reference. Not a link. <<

A pointer is the most genreal term; basically it is a machine address that can have math done on it. The term link refers to pointers in LISP and operating systems which have restrictions as to where and to what they can point. A reference is the most restricted construct because it has a type, no math allowed and has to exist, not in the hardware, but the data. How it works is implementation defined. Picky picky, picky, but spend a decade writting ANSI Standards and you get that way.

>> .. the notion of programming language concepts and relational
database concepts belonging to some mutually exclusive vernacular, which is ridiculous. <<

Actually, it is Dr. Codd <g>. He has some papers about the differences in concepts and terminology and why you need two very different mental models to do algorithms and to handle data. Hence, tuples and rows, not "records"; attributes and columns, not "fields"; tables and relations, not "files"; there is nothing in a file system that resembles a schema.

>> Index as a programming concept (for arrays) in no way is similar to
the concept of index as a relational database concept. <<

That is because there are *no* indexes in the relational model, or in Standard SQL. Those are vendor extensions and not the only possible ones -- hashing (Teradata), bit vectors (Kx, Nucleus, Sybase IQ, etc. ) and so forth. Received on Wed Sep 25 2002 - 20:03:58 CEST

Original text of this message