---------- Forwarded message ----------
From: Arthur Fuller <fuller.artful@stripped>
Date: Fri, Feb 13, 2009 at 3:36 PM
Subject: Re: Codd's rule 8 (physical data idependence)
To: Yusuf Khan <yousuf.khaan@stripped>
Sad to say, you cannot casually switch engines and hope that everything
shall continue to work. This is after all the real world, in which politics
is the art of compromise. Some things port, some don't. C'est la vie, c'est
l'amour, c'est la mort. Life is tough. Not one single vendor has made (nor
even bothered) to make this possible. I don't vilify them for this. One sets
one's priorities, and in the case of numerous for-profit vendors, one
introduces enhancements which are guaranteed to be non-portable. In other
words, "vendor lock-in". In the real world, the late Dr. Codd (much as I
admire him) doesn't hold much weight. The marketeers win out. I hate that
this has happened and I am one of the biggest champions of Dr. Codd and Mr.
Date, but in the face of marketeering that means squat.
In the case of MySQL, this is decidedly a double-edged sword. The MySQL team
has introduced several very powerful extensions to the language, none of
which are accepted in the SQL 92+ syntax. That doesn't dismiss their power.
It just means that the syntax is not portable to other SQL implementations.
(Case in point -- the multiple-insert syntax in MySQL has resulted in an
equivalent syntax in MS SQL 2008, so now both of them break the mold.)
On Fri, Feb 13, 2009 at 6:28 AM, Yusuf Khan <yousuf.khaan@stripped>wrote:
> Hello all
> Does MySQL 5 conform to Codd's rule 8, i.e. physical data independence,
> which says that:
> *Applications should not be logicaly impaired when the physical storage or
> access ethods change.*
> Any help would be greatly appreciated.