Lists Home |
Date Index |
"Thomas B. Passin" <firstname.lastname@example.org> writes:
> [Bullard, Claude L (Len)]
> > And it leads back to the question: can a PSVI or TAI/ATI
> > be reduced to a transform? Reading the www-tag discussions,
> > it seems that XML may become even more complicated and
> > less flexible. Type augmentation seems to be an application
> > concern, not an XML Infoset concern. The only way I know to
> > keep those distinct is to regard augmentation as transformation.
> For a given xml document, there are many schemas that could fit it. Among
> many possible choices with xml schema, you can choose to use global elements
> vs complex types, just as one example.
> Therefore, there is no complex type that uniquely characterizes an element.
> Perhaps the same is true of simple types (except for basic strings), since
> you can always find other patterns or restrictions that would produce the
> same result.
> If this is the case, then it seems to me that type information, and
> especially complex type information, cannot be said to be a property of the
> document or element in itself. Instead, it is a joint matter between the
> document and the schema. If such type information is forced into the
> document representation, it could greatly reduce the inherent flexibility
> of an xml document to be processed different ways for different purposes.
> It seems to be, then, that type decoration should remain separate from basic
> infoset information, which should characterize or model the document based
> on its inherent construction.
Agree with all the above, cogently argued.
> This is another reason - add it to the others - to strive to keep PSVI
> separate from the basic infoset of a document.
Just checking in case I'm missing something here -- I believe that in
the terms described above the W3C XML Schema REC does precisely this
-- it is _very_ careful not to change the input (basic) infoset other
than by addition, so that with the exception of defaulted attributes
(an exception _mandated_ by our charter to reconstruct DTD
functionality, I might add), if a processor looks only at 'basic'
infoset properties inn a PSVI, they will see exactly the 'basic'
infoset input to schema processing.
Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
W3C Fellow 1999--2002, part-time member of W3C Team
2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: email@example.com
[mail really from me _always_ has this .sig -- mail without it is forged spam]