Re: Is Oracle deliberately difficult?

On Thu, 07 Sep 2000 13:22:47 GMT, Joe Maloney <jrpm_at_my-deja.com> wrote:

>TJI, dBase was relational in the way it dealt with data. It just wasn't
>SQL.
Duh? Wasn't aware there was a "relational way" of dealing with data. It either does Codd's 12 rules or it doesn't. Period. How it manipulates the data on disk to do it is what defines how efficient it is at doing relational. Dbase was OK for basic data manipulation, but when it came to "traditional" relational features like changing table definitions on the fly, it used the "brute force" approach: copy into new structure. Not a problem with the volumes at the time, but totally unthinkable when we are talking Tbytes. Or even Gbytes.

>Somewhere in the mid-80's relational and SQL became synonomous to
>most, but actually they are different.

Sure.

>SQL could (and even was, in
>ways) applied to non-relational databases and file sets (I remember
>EQL, which was similar in many ways to SQL, agains IMS databases in the
>late 70's.)
>

And others. Univac's DMS1100 had a query processor that was almost relational in its approach. Prime's Codasyl dbms was actually accessible by a common language that could go into flat files, Codasyl, ISAM, whatever, with the same syntax.

But relational started when Ted Codd put out what makes one. At the time, the only RDBMS in the market that satisfied all of the requisites without "brute force" approaches was ORACLE. Ingres was close, but failed on what should have been simple things.

Cheers
Nuno Souto
nsouto_at_nsw.bigpond.net.au.nospam
http://www.users.bigpond.net.au/the_Den/index.html Received on Fri Sep 08 2000 - 07:03:06 CDT