Re: ISO Specification and Schema

Daniel Cazzulino wrote:

> It would be great if a (non-normative) W3C XML Schema was available 
> for the ISO spec. Not all devs (including myself) are familiar with 
> RelaxNG. For an implementer (like me), it would also make it easier to 
> validate the schema, as .NET doesn't come with a RelaxNG validator.
> Also, a Schematron-only schema would be useful to make an 
> implementation self-"hosted", in the sense that it could validate 
> schematron schemas by itself, without depending on external 
> RelaxNG/XSD validation (in this case, it wouldn't be much of an issue 
> the lack of an XSD schema).
>  
> /kzu

I agree that there should be Schematron schema for Schematron, just that 
it is
better not part of the ISO standard.*** I'll try to make one in the 
coming week.

I also should be putting out an HTML version of the draft standard: 
Elliotte
Rusty Harold will not put links to PDF on his website or even read PDFs,
and I would like to have a link on www.ibiblio.org/xml because it is a 
great site! 
(Given that he uses Macs, which have postscript built-in, I don't quite 
get his
problem, actually. )

Cheers
Rick Jelliffe

***I used to have a Schematron schema for Schematron, but we deleted to 
our surprise because:

 * The idea behind ISO DSDL multiple little languages, used together.

 * Compact-form grammars display the most amount of information per 
page, for simple
constraints. Element-based syntaxes, even Schematron to some extent, are 
not suitable
for reading.

 * I really wanted to keep the page size down. There are only 20 
substantive pages,
I think, now. Large specs get implementers lost. They give up when the 
information
is there, or they ferret around frustrated for information that is not 
there. A small
spec lets the implementer quickly figure out if what they need is there, 
or if they
need to check other implementations or ask around. If ISO allowed reference
implementations, I would have put one in in.

* There is an ISO rule against expressing the same constraint twice in 
normative
sections, which opens the doors to inconsistency. So I decided it might 
be too
brain-twisting to think of wordings for the assertions that did not 
interfere
with the normal text of the standard; I certainly did not want to have 
assertion
text at all. Once solution would have been to have made the Schematron
schema for Schematron as non-normative text, but then I would have needed
the RELAX NG or DTD anyway.

* XML Schemas leaves such little room in people's brains that Schematron 
must
appear very small and friendly, otherwise it will not fit. Just now 
people are
starting to figure out that when they want to serialize graph-structured 
data
into a tree, the grammar-based schema languages are properly incabable of
expressing many kinds of important constraint.

* As long as the formalities are correct (the predicate logic, the 
schemas, the
form of the query language binding),  the ISO Standard will do its job. 
It does
not have to be a tutorial: a standard is the prime resource for 
verifying our
understanding, not the prime entry point to getting understanding IYKWIM.



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com