Re: Principle of Orthogonal Design

From: Brian Selzer <brian_at_selzer-software.com>
Date: Thu, 07 Feb 2008 19:45:55 GMT
Message-ID: <TdJqj.7976$Ch6.3459_at_newssvr11.news.prodigy.net>


"JOG" <jog_at_cs.nott.ac.uk> wrote in message news:8c3135a5-8ef7-4da7-a99e-09d5cc83466a_at_d70g2000hsb.googlegroups.com...
> On Feb 7, 1:54 pm, "Brian Selzer" <br..._at_selzer-software.com> wrote:

>> "JOG" <j..._at_cs.nott.ac.uk> wrote in message
>>
>> news:c8bd75c2-5a4b-4456-8646-4ed912cc578b_at_i29g2000prf.googlegroups.com...
>>
>> > On Feb 7, 11:21 am, "Brian Selzer" <br..._at_selzer-software.com> wrote:
>> >> [snip]
>> >> A join is not a constraint, is it?
>>
>> > Now that's a very interesting question. Let me be polemic and say I
>> > think it is.
>>
>> Interesting indeed.  So, for example, the join of relations
>>
>> P {A, B, C} and
>> Q{D, E, F, G}
>>
>> yields a relation with the heading,
>>
>> {A, B, C, D, E, F, G}.
>>
>> Can you explain what is constrained by P JOIN Q?
>

> The resulting set of tuples

That it is the result that is constrained is significant here. I perceive a distinct difference between a constraint which limits what /can be/ and a query which determines whether something /is/. Just because something /can be/ doesn't mean that it /is/ in the same way that just because something /isn't/ doesn't mean that it /can't be/. Whether something /can be/ is most easily expressed using a universally quantified sentence, whereas whether something /is/ is most easily expressed using an existentially quantitifed sentence. (Most easily for me, at least, with positive predicates.)

> lets call it R is. So again, your example
> in set builder notation:
> R = { {A:a,...,G:g) | (A:a,B:b,C:c) $B":(B P $B"J(B (D:d,E:e,F:f,G:g) $B":(B Q }

>

> Here we're constraining a possibly infinite set of tuples, just as if
> we were constraining a set of numbers, S, to come from the naturals
> and to be less than ten we'd have:
> S = { x | x $B":(B P $B"J(B x < 10 }
>

> Course, I guess the perspective I'm throwing about above wouldn't make
> any sense to proponents of there being a separate "header" entity,
> which I think is D&D's line (as opposed to Pascal's say). Any thoughts?
Received on Thu Feb 07 2008 - 20:45:55 CET

Original text of this message