From: Koen Holtman (Koen.Holtman@cern.ch)
Date: Mon May 22 2000 - 14:49:21 PDT
...with personal views on RDF, the W3C, interoperability, the web, etc
On Thu, 18 May 2000, Rohit Khare wrote:
> Wireless Web Fight Gets Catty
> by David Sims
Mobility Focus of WWW Confab
by David Sims
This article mentions that 'push' is back. I think not! But you
gotta admire Wired for trying...
Some of my own observations follow below. This is a fairly personal
review, and the reader should note that I do web stuff mainly as a hobby
and justify www conferences to myself as a form of vacation.
I was there for the main programme and for developer's day.
Compared to last year in Toronto, there was less free food, there were
fewer parties, and the weather was worse from Tuesday afternoon on. The
food situation: no included lunches, and coffee breaks in which only one
type of pastry was served, no other food. I would have expected more food
for the hefty conference fee I paid, especially as I paid it out of my own
pocket. Evening receptions were pretty good though. By the way, speaking
from my experience as a Dutch national: the Dutch national kitchen (in so
far as we have one) cannot be blamed for this, a traditional Dutch coffee
gathering is supposed to include at least a few different types of cookies
and pastry. So I don't know exactly what is to blame for this.
This year's keynotes were mostly about mobile applications. The
conference committee made a big mistake though IMO in having each keynote
last for 1.5 hours, with a single speaker. I mostly dropped out of
attending keynotes after the first one. Strange thing: no keynote by Tim
BL this year.
To my taste there could have been less parallelism in the parallel
tracks. We had 7 parallel tracks going on, so I found myself
track-hopping a lot. There was so much parallelism that likely,
different groups of people have walked away from this conference with
completely different ideas of what the current and next big things are.
Reading some parts of the articles at the URLs above I have a hard
time imagining I went to the same conference.
According to WWW9, the next big thing is RDF. Last year's next big thing,
XML, was reduced this year to being one of these technology things that
support RDF. Prior to WWW9 I have had little time to learn anything about
RDF, but according to what I learned at this conference it is 1) a
research-y development project in the W3C and 2) a standard format for
encoding knowledge as a directed graph in which the edges are binary
predicates, where each predicate is identified by an URI. All nodes with
incoming edges are also URIs. The cool thing about RDF is, so I learn,
that it enables 'the semantic web'. Also RDF is like mathematics so it is
a universal culture-independent format. I asked Tim to tell me what kind
of cool mathematical operations he thought would happen on these RDF
structures once they exist as content in the web, but he told me he did
not want to give me examples, because then everybody would just think RDF
was for that whereas it really was for a lot more things. Also there was
no killer app, at least not one he wanted to tell me about. So,
personally, from my IETF background I this left me thinking 'no smooth
upgrade path', and from my European CS background I was thinking 'AI
handwaving alert!'. Mind you RDF could be very cool if it works and gets
deployed but I am not holding my breath.
Also, though RDF defines a way to represent sets of binary predicates, it
does not define much semantics beyond that. All predicates on the web are
supposed to live together in a big 'AND' formulae, that is one piece of
the extra semantics, and predicates can also be assigned trust levels. A
lot of extra semantics is needed if RDF is to become useful, but I got the
impression that the W3C does not want to touch defining any of it. It is
the usual W3C party line: 1) everybody can define anything by creating an
URI, isn't that great? 2) what do you mean you are worried about
interoperability if 100 people create URIs for often slightly different
notions of 'the color red'? We have this framework in which you can
formally express that all 100 are somewhat equivalent!
The same party line was still present in the CC/PP effort (content
adaptation). In the end a big chunk of the interoperability problem is
not addressed. So here is an obvious area in which commercial players can
try to create de-facto standards and leverage lock-in strategies off them.
Though I don't believe this was the secret agenda all along. Instead I
believe that many W3C people managed to genuinely convince themselves that
they do not need to address this part of the interoperability equation
(which is arguably the hardest part) when working to extend the web.
Sigh. In any case I kind of committed myself during the conference to
look seriously at using the IETF media feature tag registry to promote
interoperability between vocabularies in content negotiation and
adaptation schemes. Yet another crusade to sink my time in; just what I
Outside RDF, in the nearer term, I got the impression from various talks
and panels (mostly those with some W3C involvement) that the benefits of
of HTML 4.0/XHTML are just around the corner. I'm not so sure myself, but
one can hope. A panel on web acessibility was fairly optimistic overall
on getting there, partly due to some new US legislation I believe. In
the panel on internationalisation (in which I participated) we ended up
being less optimistic.
Mobile web-like applications are coming, though how closely they will
resemble the web, and how closely they will be integrated with it, is
uncertain. It is also uncertain whether the 'web author in the street'
will either be willing or able to write content targeted at/suitable for
mobile users. So whether the mobile wap/web will ever have zillions of
sites and pages remains uncertain.
The level of the refereed papers was good, at least that was the
impression I got from the few paper sessions I attended. There was a
huge number of posters (75), many of them good.
All in all I am glad I went. WWW10 is in Hong Kong, not yet sure whether
I will want to fly out to that one.
This archive was generated by hypermail 2b29 : Mon May 22 2000 - 15:57:07 PDT