Re: Check out the DCD submission

Peter Murray-Rust (peter@ursus.demon.co.uk)
Tue, 11 Aug 1998 20:17:58


At 09:25 11/08/98 -0700, Tim Bray wrote:
>At 09:21 AM 8/11/98, Peter Murray-Rust wrote:
>>I am primarily concerned with the basic datatypes section 4. I want to use
>>these (or something like them) in CML, healthcare and so on. I shall
>>extract this subset of the document and re-use the concepts in JUMBO2.
>
>Not now! Everyone agrees that we need datatypes. I expect some cheerful
>& constructive bloodletting in the SIG & WG as to which set of datatypes
>we need. Anybody who invests effort in that set at this time is spinning
>wheels.

It's not a problem :-). The main effort is in getting the validation,
presentation, etc done. To add new datatypes or syntax is relatively
straightforward. And I'm not doing the whole lot, either - just the
commonest ones. And I wouldn't go near the COBOL stuff.

>
>On the other hand, bearing what you say in mind, I think there may be
>a sane case to be made for decoupling the task of identifying the set
>of datatypes from that of defining the rest of the schema mechanics;
>however we end up specifying the datatypes, the heavy typechecking
>coding will be about the same, and it might be nice to give that a
>head-start.
>
You yourself came up with XML-TYPE or whatever it was called - about 10
webyears ago. IMO that was nearly a no-brainer and orthogonal to any other
effort - it would have helped other activities since.

And the arguments are unlikely to stray beyond what the precise definition
of some weird type is... I can live without min and max in that spec if it
helps.

>>I note that max and min have changed from being content to attributes.
>
>No! Read the early part of the spec. Per RDF, these things are
>*properties*, which can be given either as elements or attributes.

I admit I haven't given it very much detailed attention. But given that
there are ElementDefs and AttributeDefs and that the latter produce
attributes... But I will read it more carefully. NNTR

> -Tim

Peter Murray-Rust, Director Virtual School of Molecular Sciences, domestic
net connection
VSMS http://www.nottingham.ac.uk/vsms, Virtual Hyperglossary
http://www.venus.co.uk/vhg