Re: Guessing?

"paul c" <toledobysea_at_ac.ooyah> wrote in message news:TSe%j.302043$pM4.279613_at_pd7urf1no...
> Brian Selzer wrote:
>> "paul c" <toledobysea_at_ac.ooyah> wrote in message
>> news:V62%j.172211$Cj7.105689_at_pd7urf2no...
>>> Brian Selzer wrote:
> ...
>> I think you're misapplying Darwen's use of the word 'natural.' That one
>> design is more asthetically pleasing than another says nothing whatsoever
>> about whether the underlying logic is consistent or not. ...
>
>
> If you're right and he didn't have aesthetics in mind, he could have not
> said 'natural'. If you're wrong, maybe he had aesthetics in mind, which
> would seem mystical to me.
>
>
> The choice here is
>> between a design based upon the disjunction of two unary predicates and
>> one based upon a single disjunctive predicate. The former has one
>> relation for each unary predicate; the latter has everything stuffed into
>> one relation but requires the addition of a flag attribute in order to
>> force exclusivity
>> (i.e. x /\ ~y | ~x /\ y | x /\ y).
>> ...
>
>
> Yes, and when you ask the dbms 'is x true?' (in the union), you don't need
> to qualify by asking 'is x true and y untrue or is x true and y true?'.
> Asking 'is x true?' is enough. Same goes for any expression that joins
> the query's relation with the view. That is the character ('character'
> sounds anthropomorphic but I don't mean it that way, wanted to say
> 'nature') of the operation that forms the union. Your 'flag attribute' is
> effectively redundant and 'or' has in effect been removed from the
> predicate.
>

But I wasn't talking about union: I was talking about strong orthogonality.

You don't need to /ask/ it, but you do need to /store/ it. A disjunctive predicate, a \/ b, can be satisfied if a only or if b only or if both a and b are satisfied. That a tuple exists implies only that at least one of a or b is satisfied by the tuple. It doesn't specify which. That's what the flag is for. The flag indicates whether only a is satisfied, whether only b is satisfied or whether both a and b are statisfied.

>
> (as maybe not everybody knows, the algebraic answer to 'is x true?' is not
> necessarily 'yes', more conventionally it is a relation containing a tuple
> that stands for the 'x' proposition although Darwen has shown that an
> algebraic query could be written to produce 'yes' or 'true'.)
>
>
>
>> Stuffing everything into one relation may at first glance appear to
>> simplify when in fact it actually complicates. Consider, for example,
>> the following equivalent schemata:
>>
>> P{X}
>> Q{X}
>> R{X, Y} R[X] references P[X]
>>
>> as opposed to
>>
>> S{X, F} F IN {1, 2, 3}
>> T{X, Y} T[X] references (S WHERE F = 1)[X]
>>
>> Note that the foreign key in the former becomes much more complex in the
>> latter. I think, therefore, that the KISS principle would argue that
>> disjunctive predicates for relations should be avoided whenever possible.
>
>
>
> That seems to be part of an argument about something else.
>

Indeed. It is not about unions but rather about strong orthogonality, which is what I thought we were discussing. Received on Wed May 28 2008 - 19:49:06 CEST