W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2000

Re: RDF An opportunity

From: Sankar Virdhagriswaran <sv@hunchuen.crystaliz.com>
Date: Fri, 6 Oct 2000 11:36:04 -0400
Message-ID: <00dc01c02fab$23a4f460$4975ef8c@crystaliz.com>
To: "Craig Pugsley" <craig.pugsley@mimesweeper.com>, "'Betsy Skillings'" <BSKILLINGS@llbean.com>
Cc: <www-rdf-interest@w3.org>
This is a soap box. Please ignore if you don't like reading soapbox
pronouncements. Apologies to people who I may be offending.

I don't know how many times I have said the same thing over and over again.
But, it is worth repeating.

When this email list was started, one of the goals was to 'simplify RDF'. To
a lot of RDF experts, this meant one of two things : simplify the syntax and
simplify the RDF model so that it can layer more and more sophisticated
modeling. A third thing got added to the discussion later: we have bad
syntax, so let us put a GUI to hide the syntax.

In my opinion, all of these *will not* make RDF simple to use.

What will make RDF simple to use is a brain damaged simple model that
majority of the folks who have some modeling experience can understand.
Sorry, first order predicate logic is not it. Sorry, an object system that
is built on top of first order predicate logic is not it. And, sorry, having
a UI on top of all this not it either.

Just look at the existing markets to see why this is the case. Can you tell
me the market size of entity-relationship modeling tools vs. object modeling
(as in UML) vs. knowledge modeling tools. The last is close to $0. People go
from what they know. The Web spread because one had latex documents and Word
documents one could convert in a jiffy and put it on a site. Initially, the
effort to understand and use the Web was close to zero for most authors.

Knowledge acquisition and modeling are infinitely more difficult tasks. The
largest number of models out there are entity-relationship models. Let us
start there.

I don't have sympathy with the arguement that says entity-relationship
models *can be* represented in RDF. That is not the point. Imagine this
analogy. When the web started out, if some one came in said for presentation
use XSL or XHTML and for data use XML, how many people would have done this.
Not only did very few people have documents that could mapped into these
separte 'renderings' (as compared to Latex to HTML or Word to HTML), but
even fewer understood the separation of content from presentation (notice
what happened to SGML which had made that point for a long time).

We are doing some thing very analogus with RDF. RDF is simply too
complicated, too general, too abstract for most people to grasp and use.

Sorry, I feel that the world would be a better place if standards such as
RDF are used by the 'people' rather than a standard that only a few high
powered logicians or knowledge engineers can understand and use.

Sankar
Received on Friday, 6 October 2000 11:33:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:44 GMT