[
Lists Home |
Date Index |
Thread Index
]
> Jeff Lowery scripsit:
>
> > I guess the point I'm trying to make is that we seem to narrow our
> > definitions of universal types to only those that have validatable
> > membership and universal representation. There are a lot
> of well-understood
> > types they fail both criteria.
> >
> > Prime numbers.
>
> Say what? Every prime number has a unique lexical representation,
> and there is an effective test for telling primes from non-primes.
Sorry, I must be abusing terms again. By 'lexical' I meant something like a
number being expressible as arabic or roman numerals. If that's the wrong
term, give me another.
> That looks like universal representation and validatable
> membership to me.
Not practically, I didn't think. If I'm wrong, change it to the type whose
members are the result of multiplying two primes. I hope that one's still
tough. Again, abuse of terminology: computable, yes, but not in a practical
sense.
I'm guessing that there are types whose members can only be reasoned by
induction, but I could be wrong. It's the impracticality of validation of
some members that I'm trying to express. And that may vary depending on
local resources.
> Whether it's important to any particular user to actually validate is
> another concern. If so, the receiver just specifies a local
> scheme that
> specifies xsd:integer rather than jl:prime as the type.
Yeah, but he's not really validating the value to the extent necessary to
ensure his computations are giving correct results.
> Now if you wanted a type without either, consider "beautiful
> thoughts".
Do you have a rigorous definition for those?
>
> --
> John Cowan jcowan@reutershealth.com www.ccil.org/~cowan
> www.reutershealth.com
> "If I have not seen as far as others, it is because giants
> were standing
> on my shoulders."
> --Hal Abelson
>
|
