Re: [xml-dev] There is a meaning, but it's not in the data alone

[John Cowan <cowan@mercury.ccil.org>]

Bullard, Claude L (Len) scripsit:

> So the RNG has not even a hint of interpretive semantics 
> and doesn't use a DOCTYPE.  I suppose a document could 
> theoretically stuff a PI in there to point to the 
> RNG if that were handy.

Yes.

>  A "validator" as you used 
> the word below is some process or person that needs 
> to perform validation using RNG, yes?

Yes.

> Now I understand what you meant by 
> "DTD on steroids"; RNG does what most of us do with 
> DTDs most of the time.

Indeed.  But I was also talking about its syntax (specifically
the non-XML syntax that is generally used for schema authoring)
and the overall feel and capabilities.

Compare http://www.ccil.org/~cowan/XML/ibtwsh6.dtd
and http://www.ccil.org/~cowan/XML/ibtwsh6.rnx

> Now why would RNG want to decorate with 
> types?   How would that be different from what Schema 
> does?  

It's one of the many kinds of things that Schema does.
But it's natural for RNG to do it: the program parses
the XML against a BNF-like description, and there's no
reason it can't generate a set of indications of just
which BNF rules were applied and how.  However, the
grammar is allowed to be ambiguous, which doesn't matter
for validation, but does matter for type assignment:
we can only assign types insofar as the rules are
non-overlapping.

-- 
John Cowan           http://www.ccil.org/~cowan              cowan@ccil.org
Please leave your values        |       Check your assumptions.  In fact,
   at the front desk.           |          check your assumptions at the door.
     --sign in Paris hotel      |            --Miles Vorkosigan