OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index]
Re: [xml-dev] Recent allegations about me

Rick Marshall said:

> They're not really standards.
> Having the imprimatur of a standards organisation no longer means you
> are reading a standard. Why? Well now there's 2 document storage
> standards which is an oxymoron. There is a standard or there isn't a
> standard.

At ISO SC34 there are currently 2 to 5 different standards for
grammar-based schemas (depending on how you count them): DTDs,
Architectural Forms, RELAX NG, RELAX NG compact syntax, and Namespace
Aware DTDs.  Plus 1 if you count W3C Schemas too. Overlap is not
necessarily a problem, where there are historical or technical or
functional or procedural reasons.

At SC34 there was, in the past, a view that it should be in the business
of making "enabling standards" rather than proscriptive ones. Richer
infrastructure rather than monocultures.

A standard is essentially an agreement, in ISO terms. When there are two
groups who cannot agree, sometimes having two agreements is the best that
can be done. At least it reduces the field and brings out the requirements
into the open. If one group could try to, but does not, block the other
with procedural ploys; the other group should reciprocate, if they are
interested in respect and fair play.

Any extensible standard where conformance is defined largely by validation
against a schema, and which allows arbitrary binary media and scripts
cannot guarantee full interoperability between applications, whether it is
brand A or brand B. So governments who want level-playing field and
interoperability need to do more than just say "Just use any ISO standard"
even if there is only one standard. They will need policy to say "Here is
when you use PDF, here is when you use HTML, here is when you use etc." 
They will need profiles to say "When you use this format, here are the
allowed values and foreign namespaces you can use". Schematron is probably
good for specifying profiles. And when it comes to application
conformance, they need test suites and testing bodies. This is necessary
whether ISO has 10000 standards or 10001 on its books.

Maybe the smart issue isn't "How do we force people to use the standard we
like" or "How do we force vendors to adopt their competitors' native
formats as their own default formats?" but "How do we make it easy for
people to migrate and open up on their own timetable and according to
their own priorities?"

Think about XML and its character encodings system. When the proposal came
up to allow multiple encodings, there was dismay from the Unicode people
who wanted UTF-* only mandated, not legacy encodings. But XML has provided
an enormous on-ramp for Unicode, a Trojan horse, because it takes the
needs of legacy data very seriously. It says, in effect, that if UTF-* is
in fact technically superior and what the market is ready for, that should
be their call on their time. Now Perl has adopted Unicode, finally
prompted by the need to support XML. And now, after 10 years, we probably
*could* restrict XML to coded versions of Unicode (UTF-* plus any national
CJK encodings of Unicode) without people losing sleep.

Rick Jelliffe

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index]

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 1993-2007 XML.org. This site is hosted by OASIS