On 4/9/2012 7:17 AM, Kingsley Idehen wrote:
> The desire of RDBMS vendors to lock-in customers ultimately rammed ODBC
> down a predictable technology cul-de-sac. (01)
Please note the timing of the ANSI\SPARC conceptual schema: 1978.
The R & D that led up to that occurred during the height of the
"database wars" among the firmly established IMS (hierarchical),
the proposed standard CODASYL DBTG (network), and the new upstart
relational DBs. (02)
Even within IBM, the big cash cow was IMS, and there was a huge
internal battle between the firmly entrenched IMS and the research
guys. Even today, there are a lot of customers who still run
their old IMS DBs. (03)
The goal of supporting interoperability among those systems was
a very high priority, and it involved exactly the same kinds of
problems that we are faced with today. And the work continued
with the ISO TR9007 in 1987. (04)
> Thus, today we have better context and technology for making the
> ANSI/SPARC conceptual schema idea of yore a contemporary reality. (05)
Yes, but it could have and should have been much better. We're
just beginning to dig ourselves out of some disastrous decisions
from the late 1990s. (06)
I like the vision that Tim B-L presented in his book. But the
Semantic Web was hijacked by two groups who were pursuing the
worst kind of *technology driven* design: (07)
1. The XML crowd, whose goal was *XML everywhere* despite the
fact that the only practical experience with *ML languages
was for formatting documents and for linking HTML. There was
*zero* understanding of how to design XML-based languages. (08)
2. The decidability crowd, who had been proving theorems about
decidability since the 1930s and about computational complexity
since the 1970s. But practical programmers ignored them. Their
goal was *revenge* -- ram decidability down the throats of all
those unwashed programmers. (09)
Please note that neither of these two groups had any commercial
successes of any kind. Netscape was one of the pioneers in the
promotion of XML, and they made the disastrous decision to build
version 5.0 on top of XML. And it drove them into bankruptcy.
They donated the pieces to the Mozilla foundation, which finally
made the browser successful (because it's free). (010)
Today, I happily use the XML-based products from Mozilla,
Open Office, and now Libre Office. But they are still using
XML for the sweet spot: formatting documents. That's good. (011)
For data formatting, there is some use of XML, but JSON is
rapidly replacing it. And JSON, by the way, was designed by
Netscape as part of JavaScript. If they had stuck with their
winner, they might still be in business today. (012)
As for decidability, I strongly urge every programmer to take
a course in decidability and computational complexity. But
programmers have discovered very good methods of controlling
complexity by using structured design methods. Similar kinds
of structured design can be used to control complexity in
knowledge representation. People in AI have over 50 years
of research and practice in doing that. (013)
As for description logics, the most successful were LOOM and
PowerLOOM. But their designer, Bob MacGregor, said very
clearly: customers always ask for more expressive power;
they *never* ask for decidability. And he was right. (014)
In summary, the vision for the SW was good. But that vision was
destroyed by the worst kind of technology-driven design. (015)
John (016)
_________________________________________________________________
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 (017)
|