[ACTION-410: Review RELAX NG Schema]

I've reviewed the RELAX NG schemas supplied by Makoto and I see no
obvious issues with the schemas themselves. The challenge, as mentioned
by Scott earlier, is that 

x(XML Schema) < x(RELAX NG) < x(XML Signature specification)

where x() is the capability for exactness as to what the format should
be. Ideally, it would be nice if one could just derive the RELAX NG
schema from the XML Schema but because the exactness of XML Schema is
less than RELAX NG, there is the hindrance that the result will not be
as exacting wrt closeness to the actual prose specification as it could
be.

I haven't gone into a full scale comparison of Makoto's schemas and the
XML Signature specification (and I don't have time to) but I am
certainly glad to see Makoto's work. Indeed, I would suggest that
ideally (if I had the time, which I don't) it might be a good idea,
noting the greater exactness of RELAX NG, to specify, non-normatively,
the XML Signature schema in RELAX NG and use that to automatically
derive the normative XML Schema.

Thanks Makoto for your work in this. I'm particularly interested in it
because of other work I've done in XML Signature sub-classing and RELAX
NG (and also OPC). Please keep us informed. Btw, would it be possible to
supply the XML, rather than the compact, version of the RELAX NG schemas
in future updates?

Ed

Received on Wednesday, 21 October 2009 19:56:59 UTC