Lists Home |
Date Index |
John Cowan wrote:
> W. E. Perry scripsit:
> > Quite simply, what datatype something 'is' is a corollary of that
> > being manipulated as of that type by a process. Type--and not just
> > datatype--as opposed to labelling, inheres in process and not in the
> > text or data submitted to processing.
> This view of typing is common enough, but IMHO goes beyond necessity.
Goes beyond necessity for what?
> The RNG view of a datatype is that it is simply a pair of boolean
> functions on strings: isFoo(string) is true iff string is a syntactically
> sound representation of the type Foo, and sameFoo(string1, string2) is
> true iff string1 and string2 are interchangeable representations of Foo.
> No notion of "objects of type Foo" is or need be introduced.
I am not sure what you are getting at. One can use the phrase "objects of
type Foo" as a synonym for "the class of x for which isFoo(x) is true".
I think what Walter is getting at when he is using the term "process" is
that there are two processes at each end of the wire and for which the
semantics of xxx is represented as a constraint on how the processes are
encoded. That is to say, the semantics of xxx represents the set of possible
worlds, shall we say processes, which conform to the semantic specification
of xxx. (Hopefully I am at least close to what Walter means here.)
> And in fact this is almost the whole work that XSD can do with types
> as well. About the only remaining notion is a second-order function
> which asks of two types Foo and Bar whether every representation of Foo
> is a representation of Bar as well.
I assume you are saying that XSD can do -without- types, and to the extent
that XSD is used as a language to express -syntactic constraints- on unicode
strings, this is certainly true. The point of semantics is that now that we
have a piece of XML that let's say corresponds to some RNG pattern or XSD
type _now what_. How can we specify how different programs i.e. processes
might use this piece of XML in an interoperable fashion. For example suppose
Penicillin 500mg p.o. q.i.d. x 3 weeks.
(assume an XML encoding of that since we are xml-dev here)
I want to be very precise about what happens when this message is sent to
the pharmacy -- at least as precise as the current situation when I write
this down on a slip of paper and transmit the message via sneakernet.