[sc34wg3] Newcomb on "Where WG3 should be going"

[michel]

Here is my contribution to the current discussion about directions for
WG3.

I would first like to observe that I think the time for
Topic Maps is before us, not behind us. The market is only now
starting to be ready to implement what we have been thinking about
since a long time, and there are encouraging signs of new
interest. However, the way things are now may be a bit different than
when we started. Back then, Topic Maps was originally designed to
provide navigational tools for electronic publishing, and it was
conceived as a generic approach that encompasses what is traditionally
achieved with indexes, thesaurii, library catalogs,
cross-references. Topic Maps have mainly be used as a way to organize
web sites putting topics at the center. The applications developed for
topic maps resemble databases with pre-defined properties that have
given rise to various products and APIs. In addition, Topic Maps
provide a general method to address things by their subjects,
rather than by their identifier, as traditional databases do.

There are several reasons why things have evolved.  A main one is that
information is now primarily available from the Web, and finding aids
have become the way to access information. Pervasive search
technologies, based on proprietary algorithms, drive the show and the
market, and have oriented the way we search and find according to a
set of often implicit requirements, that aim at maximizing the
profitability of the companies providing such technologies. Users
benefit from those technologies, but there are important requirements
that are not completely covered. I think it's in the nature of a
standard to provide features that are in the common interest, in
addition to open opportunities for companies to build profitable
business models around them. The issues is in distinguishing the near
term from the long term. It is my perception that both Patrick and
Steve are looking for solutions that could be applicable in the long
term. I also happen to be interested by those.

The publishing industry, after some turmoil due to the switch between
the print model and the digital model, is starting to rebounce. There
is still quality information available on the Web. It has not become a
place for generalized junk, as some may have feared. The problem that
remains is the quality of finding aids. We find things, but there may
be things we don't find, and we don't know it. The amount of
information which is missed is, by nature, unknown. Users who have
high expectations for finding quality information are suffering from
imposed ignorance. This is an area that can be improved by increasing
the "transparency of findability". Today's market-driven search
technologies aim at emprisoning users behind closed boundaries, also
called "user profiles", to provide them with targeted information, for
example advertisement. However, the mere abundance of information
available creates a change in users' expectations. Since we always get
something back, we consider that it's good enough, because we kind of
find something related to what we were looking for.

I think we have reached a point where good enough is not good enough
any more, at least for some users. For example, the publishers who
know the value of the information they are making available, want to
maximize its accessibility and its findability. The major competitive
advantage today is not only the quality of the information itself, but
the ability to find what's there, in a way which is maximally useful
for its intended audience. Lack of focus on search is one of the
issues. Therefore, the "refocusability" of information is a major
benefit to be looked for. This doesn't mean that noise should be
eliminated, first because it can't be, but also because it's the
noise, or the unexpected side effect, that often have been at the
origin of major breakthrough in human thought, paradigm shifts.

I agree with Steve that the notion of trust is what would enable us,
as a society, to increase wealth for all. I also agree with Patrick
that there is no better way than to experiment, because we need praxis
as well as theory. But we should also be aware that this situation may
not please every one. We are confronted these days with a situation
where governments are not so much interested by the public good than
by preserving the ability for the powerful to dominate economically
and politically over the masses which don't benefit from
redistribution of wealth. In that sense, we are now in a situation
where such a project has political implications, whether we want it or
not. It's similar, at least in the US, in trying to combat
human-caused global warming, which is considered a major political
dividing line rather than a direction on which consensus can be found.
Looking for consensus in the information world of today may be an
illusion, and that's not because universal consensus can not be found
that it's not worth pursuing.
There are historical lessons to be learned. When the printing
industry developed into a series of sophisticated knowledge sharing
tools, such as indexes, library catalogs, encyclopedias, at a period
known as "Enlightenment", the people involved were considered radical
thinkers and the societies were turned over. This period culminated
with the advent of the American and the French revolutions, and the
democratization of access to science and knowledge by better, more
accessible education systems. The fact that today we are finding
ourselves in a highly controversial debate which involves the
foundations of the societies we live in may not be purely
coincidental.
I am not sure what role the SC34, and ISO have to play. One thing that
can be noticed though is that when SGML was first proposed as a
standard, it went far beyond current practices of the industry and was
not considered a viable solution. It only reflected practices of some
advanced research laboratories in information technology. I see the
current discussion trend as a way to share our thoughts about where we
think things could go, as opposed to where they actually go, and
that's a very interesting moment. The experience that we have
accumulated with practicing topic maps over the years, their
advantages, and their limitations, the difference between what we
originally thought and what users actually do, is something that
should be brought in the discussion as well.

Michel
-------------------------------------
Michel Biezunski            Infoloom
New York         (718) 921-0901
Washington DC    (202) 234-1610
Cell             (917) 803-6768
mb at infoloom.com
http://www.infoloom.com
-------------------------------------