Open Standards Processes
ricko at allette.com.au
Mon Apr 27 22:24:03 BST 1998
If an author writes a book based on drafts, they are taking a gamble.
Those of us who bought "The Draft Standard C++ Library" by Plauger
will understand this. I think a responsible author will delay publication
until it is clear what the eventual standard will be--we delayed my
book "The XML and SGML Cookbook" (now on the presses) for this
reason, and I don't think that we should cry too much for authors who
put out books based on drafts.
(I talked to Sharon Adler at WWW7, and she gave me the impression
that XSL would look almost nothing like the draft. Certainly she was
insistant that it moved a lot beyond DSSSL.)
I think Jon Bosak's idea of allowing writers to monitor (if not participate
in) groups is great. A technology like XML needs to be promoted
if it will be popular.
People have been saying that the W3C process is not open.
I was invited to join the SIG as an outside expert, and not as a
spokeman for any group. This was exactly the same reason I
was invited to participate in the ISO WG4, which looks after SGML.
W3C's process is formally closed, but often, in practise slightly open.
ISO's practise, in my experience, is that it is formally open, but
in practise slightly closed. This is because you (i.e. the technical
public) choose not to participate. You do not go to ANSI meetings,
or whatever your national body is, at least not in sufficient numbers.
This means that ISO groups are largely
populated with national representatives from large companies
or specialist companies, with a smattering of independent
consultants who use ISO to make sure they really understand their
subject, and then a handful of special interest groups (American
Mathematical Association, UK SGML Users' Group).
But ultimately the ISO WG4 vote comes down to national lines. So if
US has 100 people they share 1 vote; Australia has 1 person (me),
so I control 1 vote. Now this disparity acts to stop a total control
of ISO standards by large corporations (in the recent Java standards
dispute, Microsoft and Sun had it pointed out to them that it was
not enough to win the US vote: if they did not participate in national
standards bodies they could not expect to be held in high regard
when speaking for or against standards: the attitude that seemed to
come from some representations that non-US bodies simply
rubberstamp ANSI was naive, if not repellent). But it means that, to
some extent, even ISO standards can be skewed by individuals:
this skew can only really operate when the indivuals from the
peripheral countries vote NO (we cannot force a bad technology to
be standardized, but we can block what we think is a bad one.)
Of course, I consult with other SGML and XML stakeholders here.
And anyone else is free to participate, if they have some minimum
provable expertise, according to each national standards body's
particular rules. Australia has a policy of trying to represent the
legitimate needs of New Zealand (this is a formal arrangement) and
our neighbours in the Asia/Pacific (an informal, good neighbour policy.)
(I note that the IETF procedure, to pick another "standards"-making
process, has failed over many years to address the "selective ACK"
problem in TCP/IP, which makes some kinds of Internet traffic between
peripheral countries like Australia very difficult: this problem would
have been quickly put high on the agenda if ISO procedures were
followed. People outside the US should be very suspicious of
"standards" bodies which do not have a guaranteed nation-based
review policy: we will end up with "center-periphery" technology
rather than "world wide" technology. Now that Fuji-Xerox and Keio
are participating in W3C, there is a little bit of this in W3C procedures
now, but perhaps the center has just expanded a little.)
Now that ISO WG4 has started to conduct business by email more,
it will increasingly look like the XML SIG. The W3C might be
a benevolent dictatorship, but really I think it is a mistake to believe
that ISO WG4 is much more open than W3C was: especially since
almost all the active participants in WG4 are also on the XML SIG
or even one of the Working Groups.
In both cases, they can only use whatever experts are available.
For markup languages, there are not a real lot. And I (and I am sure
many other newcomers to the XML SIG) found that many issues
being discussed are so complicated that it would hinder progress
if issues had to be re-discussed every time a new person came along:
to be actively involved requires that you study up on what has already
been done, as much as possible.
The big issue is not "openness" in my mind, but "anti-hijacking-ness".
Is there some way which a large player could, by legitimate means,
create a form of XML/SGML which, in practise, they controlled? This would
bring back proprietory data formats. This is of course the alarm
bell that rings dully with XML-data, but it also tolls for RDF. Any standard
which does not specify enough will get proprietory extensions to fill
the gaps. So "openness" in the standards-making process does not
guarantee openness in the eventual standard.
XML is not a standard. It is a specification--of a subset of SGML, with some
bells and whistles. Any standards-*proposing* process cannot be democratic
or fully open: communication, language and factionality will prevent that.
So a W3C technology, or a Sun one, is IMHO just as good a candidate
to be proposed as a standard
as a technology dreamed up by ISO committee representatives But where
ISO has an advantage is that the standards-*voting* procedure is more
open: most national standards bodies are democratic or consensual,
and the variety of their membership, and the ease with which representatives
of small companies and individuals can participate (if they bothered)
does give ISO a credibility which W3C does not have (and does not
pretend to have, as far as I have seen.)
xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)
More information about the Xml-dev