Re: 1GB Tables as Classes, or Tables as Types, and all that refuted
From: Jan Hidders <jan.hidders_at_REMOVETHIS.pandora.be>
Date: Fri, 26 Nov 2004 17:18:11 GMT
Message-ID: <nzJpd.1451$wU1.144455_at_phobos.telenet-ops.be>
>
> Then I can continue to think the same without worry :)
>
> The practical problems exist. I supose that you have readen the
> section of TTM that describes some of these practical problems, like a
> return to chasing pointers.
>
> Indeed, these were model problems :)
Date: Fri, 26 Nov 2004 17:18:11 GMT
Message-ID: <nzJpd.1451$wU1.144455_at_phobos.telenet-ops.be>
Alfredo Novoa wrote:
> On Thu, 25 Nov 2004 21:32:50 GMT, Jan Hidders
> <jan.hidders_at_REMOVETHIS.pandora.be> wrote:
>
>
>>>>I simply cannot believe that you do not see what is problematic about >>>>this type of argument. Seriously. >>> >>>So what is problematic? >> >>It's irrelevant.
>
> Then I can continue to think the same without worry :)
Sure. If you are not worried by incorrect arguments....
>>The question that must be answered is whether there are >>any practical problems if you redefine the notion of type such that it >>does include relational variables.
>
> The practical problems exist. I supose that you have readen the
> section of TTM that describes some of these practical problems, like a
> return to chasing pointers.
Yes, and as anybody who has actually done some research on OODBMSs knows that is blatant nonsense.
>>>It's well known that the DBMSs that tried to treat relational >>>variables as types were a fiasco. >> >>It is also well known that these were not technical problems.
>
> Indeed, these were model problems :)
Nope. There were also no model problems.
- Jan Hidders