On 5/26/2011 11:38 AM, Zhuk, Yefim wrote:
> From technology perspectives, I'd think of Semantic Web as a chain
> of beautiful ideas: Internet - Web Links - Linked Data - Distributed
> and Linked Knowledge. (01)
I have no quarrel with that point. And as I said many times, I was
enthusiastic about the potential for the SW back in 1998. (02)
But what bothers me is that they ignored a huge number of other *very*
beautiful ideas, which were just as important as the ones above and
which desperately needed to be integrated with the above: (03)
1. When the SW was just getting off the ground, every commercial web
site was organized around a relational database. Big ones used
Oracle or DB2, and little ones used LAMP (Linux, Apache, MySQL,
and Perl, Python, or PHP). Instead of bringing that technology
into the SW paradigm, the SW ignored it and made their technology
incompatible with it. (04)
2. The VLDB (Very Large Data Base) conferences started in 1975.
By 2000, they had a quarter century of outstanding R & D on every
beautiful idea in the SW. In addition, they had contributions
from every major branch of computer science and technology.
They had addressed *and* presented excellent solutions to the
problems of integrating RDBs, OODBs, and triple stores. And
they addressed every issue of integrating semantics with all
of the above. Unfortunately, the major DB vendors ignored
most of the VLDB research because it would make their highly
profitable software obsolete. But it was *exactly* the kind
of research that the SW needed. (05)
3. Also by 2000, AI had over 40 years of R & D on every beautiful
idea in the SW plus much, much more. The obvious notation for
triples was LISP: (A B C). But the Semantic Webbers complained
that LISP had too many parentheses. Not only did they make their
formats incompatible with all the R & D in AI, they replaced the
beautiful (A B C) notation with the ugliest and most inefficient
notation ever inflicted on poor innocent programmers. (06)
4. UML was very widely used for mainstream commercial software and
database design. The UML diagrams already represented the most
widely used aspects of OWL, and they could have been adopted and
elaborated as the basis for displaying and developing ontologies.
Today, many people are finally turning to UML as an alternative
to the SW formats. But the SW should have started with UML-style
diagrams from the beginning. (07)
5. Finally, there's a little box called logic in the SW layer cakes,
but the foundation is pure syntax. And syntax overwhelmed the
logic. When the SW was founded, KIF was widely accepted as the
major logic-based format for ontology, there were many tools
available, and there was an ISO project to standardize KIF and
CGs as a foundation for Common Logic. But the SW ignored that
just as quickly as they ignored every other beautiful idea. (08)
> The big thing for SW was the brand itself and its promotion. (09)
Yes, but a more accurate term is "hype machine". They adopted
Tim B-L for his name and reputation, but the W3C was supposed
to be a "democratic" committee. That means they made technical
decisions by voting. (010)
There were many very intelligent people in the W3C who were familiar
with every one of the five points above. But all the intelligent
people had their own special areas of interest. As a result, they
tended to go in different directions or squabble among themselves,
while the voting caused the results to degenerate to the lowest
common denominator. (011)
In summary, I agree that many of the starting ideas were beautiful,
but the committee design ended up as an incredibly ugly mess that
killed a beautiful opportunity to integrate all semantic systems. (012)
John (013)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (014)
|