Re: Database design, Keys and some other things

From: paul c <toledobythesea_at_oohay.ac>
Date: Sat, 01 Oct 2005 20:44:33 GMT
Message-ID: <RyC%e.39341$tl2.18686@pd7tw3no>

mAsterdam wrote:

> paul c wrote:
> 

>> mAsterdam wrote:
>>
>>> paul c wrote:
>>>
>>>> mAsterdam wrote:
>>>>
>>>>> ... When designing a database, we assume a closed world for
>>>>> the database under design. Aren't you sugessting we
>>>>> work within that same specific closed world when
>>>>> designing it?
>>>>
>>>>
>>>> Yes.
>>>>
>>>> Also that this makes the notion of "surrogate" meaningless within
>>>> the CWA.
>>>
>>>
>>> Well, when you are designing that big a database,
>>> you are right of course. For all other databases not, though.
>>
>>
>> That reminds me of something I read: "the purpose of metaphysics is to
>> correct metaphysics", which I think I can justify only by giving a
>> longer reply.
>>
>> I'd say the db only has to be as big as having one more attribute than
>> the smallest possible db. Would think this would include most databases.
>>
>> I'm assuming that by "smallest" we mean a db that only the empty set
>> as attribute even though there could be many differently-named
>> relations within it.

> 
> 
> I don't understand. It may be a nice exercise to explain some concepts, 
> but can such a database hold relevant facts for anyone?
> ...

I assume by "anyone" you mean "anybody except a philosopher" (or anybody except database theorists!).

Possible states - such a db might have a relation named "happiness" and another named "bankruptcy. or it might have one relation for each possible integer, ie. it could enumerate a type. Although I'm not enough of a theorist to say just how one would update it.

A better question might be "what use could it have?" Suggest that we have just put it to one use, namely, to show the pointlessness of the surrogate notion!

p Received on Sat Oct 01 2005 - 15:44:33 CDT