Re: Databases as objects

From: Thomas Gagne <tgagne_at_wide-open-west.com>
Date: Thu, 21 Dec 2006 16:33:31 -0500
Message-ID: <mMWdnfvPC4g4YBfYnZ2dnUVZ_t2tnZ2d_at_wideopenwest.com>


Cimode wrote:
> Thomas Gagne a écrit :
>
>> Why update each row individually when I can update them all at once?
>>
> Why use objects when you can use relations. RM theory simply describes
> a schema a one complex single relation.
>
>
I'm not advocating objects over relations. In fact, through the preference of procedures I'm enabling (ney--promoting) the exploitation of relational operations. What I'm trying to suggest (especially to OO programmers) is that the database is the 800 lb. gorilla they've been trying to ignore, or pretend doesn't exist by throwing a tablecloth over it called Object-Relational or any of the other mapping tools that use metadata and other various mechanism to create an illusion that OO objects are animations of relational entities.

Rather than waste effort trying to make either an extension of the other, I'm trying to demonstrate, using OO terms for OO people, that the database is an object, which should hide its data and only be used through its interface, just like they do with other objects in their models. In the same way OO programmers send messages to objects to have them do things programmers should send messages to their database to have it do things. Those things the DB might do can be fairly sophisticated and domain specific if they express themselves as procedures. Using procedures, applications can query and manipulate data in a uniform and predictable manner regardless what programming language is used or what paradigm(s) it supports.

-- 
Visit <http://blogs.instreamfinancial.com/anything.php> 
to read my rants on technology and the finance industry.
Received on Thu Dec 21 2006 - 22:33:31 CET

Original text of this message