Re: More on lists and sets

David Cressey wrote:
>
> With regard to lossless conversions, I think it's important to distinguish
> cases where the order pertains to the problem domain, and cases where the
> order pertains to the solution domain.

Yes! Nicely put.

> One offshoot of your brain dump is this odd idea: The concept of "scalars"
> in Pascal is a mapping from some other type to (a finite subset of) the
> natural numbers, 0..n
>
> Example: char maps into 0..255 or Boolean maps into 0..1.
>
> I wonder what the connection, if any, between that
> concept and this thread is.

Yes, we have at various times had threads that try to pin down just what it means to be a scalar. It seems like it should be obvious, but it certainly hasn't gone down that way.

Perhaps relational theory needs to pay more attention to sum types, instead of just focusing on product types all the time.

As far as these kinds of mappings go, it seems that category theory has a lot to say. Too bad I don't really know any category theory. (It's on the list, as are a lot of other things.)

Marshall Received on Tue Mar 28 2006 - 05:20:28 CEST