Re: Multiple-Attribute Keys and 1NF
Date: Thu, 30 Aug 2007 19:16:36 +0200
> mAsterdam wrote:
>> JOG schreef: >> >>> mAsterdam wrote: >>>> JOG wrote: >>> .... Consequently In this instance I see what one can gain by >>> encoding "Green" and "Yellow" separately in terms of flexibility, but >>> not what one would lose. Either way, I am willing for the example to >>> be interpreted in terms of such an information need. >> The example has an existence of its own, though. >> >> Your preferred interpretation requires a leap of >> faith many mediocre math-inclined formalists show >> willing to take over and over again to no avail.
> I'm not sure I follow this mAsterdam - what leap of faith am I taking?
> I'm certainly interested in cutting short any exploration of the topic
> if it lacks avail ;)
Just the faith that the problems' nature is such that a formalism can (help you) solve it. In this example I think it is clear that the problem is not in the form but in the substance. Two different kinds of information: colors as such and colors as symbols. No formal criterion can help you make that distinction.
Think of liters and kilograms - you won't add them together because your knowledge of nature makes you see the nonsense of that. This knowledge can be formalized into procedures - of course, but thats a wholly different problem from just modeling some volume and mass data about things having volume and mass properties.
There is a limit to which problems can be solved by applying formalisms. Problems of substance are beyond this limit. Received on Thu Aug 30 2007 - 19:16:36 CEST