Re: O'Reilly interview with Date
From: Kenneth Downs <knode.wants.this_at_see.sigblock>
Date: Tue, 09 Aug 2005 22:08:21 -0400
Message-Id: <934os2-pru.ln1_at_pluto.downsfam.net>
>
> You could say the same thing but use "int" instead of char(20).
>
> I just don't understand the attraction of dividing up the system into
> separate "engines". Do we have a separate int engine vs. a string
> engine? Do we have a + engine that's separate from the * engine?
> What concrete benefit does having separate engines produce?
>
Date: Tue, 09 Aug 2005 22:08:21 -0400
Message-Id: <934os2-pru.ln1_at_pluto.downsfam.net>
Marshall Spight wrote:
> Kenneth Downs wrote:
>> >> To the RM, a table is semi-structured because an item of type char(20) is >> handled as payload, an opaque stream of bytes. The type engine provides >> services such as comparisons, but the Relational engine doesn't care.
>
> You could say the same thing but use "int" instead of char(20).
>
> I just don't understand the attraction of dividing up the system into
> separate "engines". Do we have a separate int engine vs. a string
> engine? Do we have a + engine that's separate from the * engine?
> What concrete benefit does having separate engines produce?
>
There are different groups of tasks that are grouped together. Just as in your example there is a packet router that reads headers but not bodies, so is there the relational engine, which handles set operations, and the type engine, which handles interactions between actual values.
This has nothing to do with implementation, you can do a super-OO db server or a Microsoft put-it-all-in-the-OS-kernel implementation, but for general discourse it is useful to discuss the operations in their natural groupings.
-- Kenneth Downs Secure Data Software, Inc. (Ken)nneth_at_(Sec)ure(Dat)a(.com)Received on Wed Aug 10 2005 - 04:08:21 CEST