Re: Surrogate Keys: an Implementation Issue

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Fri, 28 Jul 2006 21:55:48 GMT
Message-ID: <EJvyg.28236$pu3.375638@ursa-nb00s0.nbnet.nb.ca>

paul c wrote:

> Bob Badour wrote:
>

>> paul c wrote:
>> ...
>> Not all dbms's are all that fat. Lee Fesperman's stuff over at 
>> http://www.firstsql.com/ is particularly lean. Selzer seems to want to 
>> make applications and databases fatter rather than the dbms, so I am 
>> not sure I follow you at all.

>
> I meant, for example, that it should be possible to avoid having logical
> lock manager and session manager components in an rdbms, so users would
> be unconcerned with what arbitrary locking techniques a system chooses.
>
> I'm not saying such components are 'evil', just that they hide the most
> fundamental nature of a concurrency solution by discouraging programmers
> from identifying the exact assertions that are germane. (Not talking
> about physical locking either.)

Again, I am not sure I understand what you are trying to say. A dbms must provide some facility to manage concurrency regardless whether the dbms is an rdbms. Received on Fri Jul 28 2006 - 16:55:48 CDT