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


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: Dangers of Subsetting? (was RE: Pull-based XML parsers?)

[ Lists Home | Date Index | Thread Index ]
  • From: Rick JELLIFFE <ricko@geotempo.com>
  • To: Mike.Champion@SoftwareAG-USA.com
  • Date: Sat, 11 Nov 2000 00:02:40 +0800

> Mike.Champion@SoftwareAG-USA.com wrote:
>      This nicely pulls together many of the things we disagree about
>      on this list :~)

The innovation that Mike is talking about, the ability to use the name
of a clearly specified technology developed by a known group of people
with a known set of intents, but alter it in any way that one sees fit,
might be called "lying" by some, and "passing off" by others, but most
importantly is also called "embrace and extend".  Standards,
requirements, RFCs, conventions, etc. work to channel the innovation of
large players into productive areas: XML is SGML because otherwise XML
will be whatever crapulous tags MS chooses (e.g. the ones put out by
Office 2000's HTML).  People shouldn't fool themselves that it is W3C
preventing the common man from innovating: it is MS (and anyone else
using the XML brand) agreeing to forego innovation on syntax  and to
compete on some other strategic level.

If Mike does does not believe in interoperability, that is fine, but
please don't spoil the party for the rest of us.  

Mike's examples (XML and Namespaces) have two problems.  First, XML is
an example of codifying what had been learned (for example, I wrote my
first XML-ish parser for a subset of SGML 12 years ago, and many others
have done the same over the years before XML 1.0).  Second, the problem
with the XML Namespaces spec is not that it is over-specified but that
it is incomplete: the relative URI fuss and the namespace != schema fuss
are both underspecification problems. Mike seems to have bought the hype
that XML is something fundamentally new in the SGML world: in fact, an
XML-sized subset was standard practise among data processing people who
rolled-their own systems (i.e., didn't use SGMLS or OmniMark or Balise

I am interested if Mike thinks the same thing about C? Can Richard
Stallman remove "+" from gcc (because it takes extra time to implement,
and it is redundant) and still say he has implemented C++ or C? 

Rick Jelliffe


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

Copyright 2001 XML.org. This site is hosted by OASIS