RE: [xml-dev] WSIO vs. Semantic Web

["Joshua Allen" <joshuaa@microsoft.com>]

> BUT I agree with Tim that it is ultimately Ballmer's fault that
> this is a crisis.  Nobody was smoking Web Services crack until the

Actually, I might have misread Tim, but I thought he was blaming Ballmer
for provoking him into a strong comment, not blaming Ballmer for ruining
the web and getting everyone addicted to crack.

> would have spent some time sorting out the implications of HTTP +
> XML + (pick your favorite metadata scheme) + (pick your favorite
> discovery mechanism) BEFORE promising that it would be bigger than

I completely disagree -- I think we *are* at a point where we can say
that the semantic web and webservices will be huge.  This assessment
does *not* require all of the techniques, best practices, etc. to be set
in stone.  We know that these will be huge because we can hear the
rumblings of user demand, see the degree to which technology has evolved
to make these practical, and so on.  I really don't think these things
need to be coddled, nurtured, and groomed -- they are inevitable.

> are pure dope smoke, and none of us has a very clear idea of which
> is which.

That is exactly why we have these focal points, in my opinion.  The
start of the web was not a carefully-engineered phenomenon, and neither
will these be.  At least this time around, we have even more lessons
from which to learn (and just like HTTP was informed by SMTP or NNTP, I
see plenty of evidence that things like SOAP, RDF, etc. are trying to
learn from the lessons we've gained so far).

The hype around web services and semantic web serve to clarify a vision
that people work towards.  Ballmer isn't claiming that anyone has "the
one true religion" of how to do web services; there has got to be some
degree of real-world growth of participation and feedback of best
practices.

Now, forgive me for inserting my opinions about why the perceived
mismatch in hype about web services vs. semantic web.  I think that
these are complimentary things that target different communities.  

"Web services" are ultimately more interesting to businesses.  Web
services are not a whole lot different from "the web" right now.
Vendors can have a big influence on propagating standards, etc.  And the
value proposition "what can it do for me" is graspable by most business
people, so therefore the hype value is high.

"Semantic web" in my opinion is much broader, and is interesting to
end-users.  As individuals publish metadata, it is very different from
the publishing profile of the traditional web -- Microsoft.com probably
won't provide server space for you to annotate your opinions about the
site, so your metadata will have to be stored elsewhere.  So if a user
browsing Microsoft.com wants to know what all of the annotations are how
does she know where to look for the annotations?  The point is,
"semantic web" will by necessity involve lots of decentralized and
independent work, and there is not as much power for large companies to
play a driving role.  I think the only companies that could have
significant influence on this evolution will be search engines.  Now,
because of the broad impact and grass-roots effect that semantic web
will have, I think it deserves as much hype as web services.  However, I
think that interest suffers because there is no clear value proposition
for users to grasp.  People say, "the semantic web must be cool because
TimBL says so" or "the semantic web must be cool because it will let me
use AI to figure out how many iguanas live in houses in Arizona".  But
this is not the foundation for real widespread interest.  The value
proposition that excites people should be like the one that got people
excited about the web; "You mean I can publish a document and anyone in
the world will be able to link to it?  You mean I can link to anyone
else's document and even if their server crashes, my page is fine?
WOW!"  This was a value proposition that anyone could *immediately*
participate in; nothing theoretical about it.  The semantic web is even
more exciting than that, and has just as many immediate opportunities
for people to do things that matter to them *today* rather than wait for
us to figure out "the grand unified theory of reifying everything", so I
do not know why most people think of it as "some sort of AI thing that
TimBL likes".  In fact, my personal opinion about why there is not more
activity from people writing killer apps for the semantic web is simply
that they are all got sucked into the black hole of trying to write a
knockoff version of a 30 year-old operating system to give away for free
:-P