[
Lists Home |
Date Index |
Thread Index
]
4/24/2002 12:33:29 PM, Paul Prescod <paul@prescod.net> wrote:
>
>Given these facts, I have recently tended to give the SemWeb the benefit
>of the doubt. It is probably in the same state SGML was before XML. It
>needs a simplification and some killer apps.
Perhaps, but it has generated an awful lot of ill feeling for the W3C leadership, in my
humble and personal opinion garnered from many private conversations. Whatever the
justification for that ill feeling, it is critical for SemWeb advocates to understand and
come to grips with it.
I see three general threads in the critique of the semantic web initiative:
1 - Priority: This is simply not something that most of the member companies want the
consortium to focus on. Whether or not it is true that the W3C has failed to get out in
front of the technology curve on web services, there is a persistent belief that the W3C
leadership has been twiddling with the SemWeb while WS interoperability got burned. This
resulted in the founding of the WS-I to do what the major vendors formerly relied on the
W3C to do.
2 - Progress to date: The WWW (by most accounts, again I don't claim this is reality)
emerged as a working system out of CERN in the early 1990's, and was adopted because it
solved real problems for real people (initially in the academic quasi-academic research
communities, I believe). The SemWeb by contrast has, after 4 years or so, no "killer
apps" that any but true believers find useful in their day jobs. RSS 1.0 is probably the
best know RDF application, but even it is less popular in actual use than the non-RDF
variants. I wasn't around for the early days of SGML, but as far as I know it won converts
by solving problems, not on the coolness of its vision.
3 - Probability of eventual success: this smells a LOT like some "next big things" of days
gone by that never amounted to anything. Rightly or wrongly, many people see RDF as Prolog
in XML syntax, and the vision as being disturbingly similar to the Fifth Generation
project/vision/hype of the early 1980's in which logic programming was going to put the
hackers out of business and lead Japan (the center of interest and research in this field)
to world economic domination. Other techniques/paradigms that came out of one or another
branch of the AI community over the last 25 years have had similar patterns of overwhelming
hype / underwhelming success.
OK, I'll take the flames on this :~) but I'm pretty sure this is the "conventional wisdom"
and not some idiosyncratic position of mine. I consider myself a friendly skeptic, I'd be
happy to be convinced that something like RDDL+RDF or some RDF version of a well-accepted
controlled vocabulary (e.g. SNOMED in the medical field) adds real value over what we can
do without it. So, flame me if you want, but I'd much rather that those who are saddened by
the bad reputation of the SemWeb would convince me (and most everyone else in the industry)
that the conventional wisdom is wrong.
|