[om-list] Ontology and the Web

[Mark Butler]

Here is a nice article by Nicholas Petreley entitled "Ontology and the Web" that is relevant to the One Model project:

  http://www.computerworld.com/cwi/story/0,1199,NAV47-81_STO64490,00.html

- Mark


============================================

NICHOLAS PETRELEY

Ontology and the Web
  
(October 08, 2001) The gist of last week's column was that the disorganization of information on the Web is at the heart of the growing pains we're experiencing in the Internet economy. As I mentioned last week, a Scientific American article, "The Semantic Web," by Tim Berners-Lee, James Hendler and Ora Lassila (www.sciam.com/2001/0501issue/0501berners-lee.html), proposes a way to make that information more accessible, which would empower programmers to build more intelligent software. 
The goal of the Semantic Web's supporters is to make it possible for software to find the data it needs on the Web, understand it, cross-reference it and apply it to a particular task. The means to this end are a combination of XML, Resource Description Framework (RDF) and a concept called ontologies. 

Assuming that everyone publishes data on the Web in conformance with the standards, I should be able to tell my Web-enabled handheld device to schedule an appointment with a dentist within 20 miles of home and let the computer do the rest. It should be able to interact with the data on the Web to find a doctor who belongs to my insurance plan, resolve schedule conflicts, make the appointment at both ends and compile driving directions to get there. 

Think of the Semantic Web as a project to deploy workflow management and enterprise resource planning software on a global scale, except in this case, there's no single authority over the various departments where the data is kept. On second thought, don't think of it that way; it would probably spoil your next meal. 

Obviously, I'm not optimistic. If you read last week's column, or any of my other columns dealing with XML, you already know that I'm deeply unimpressed with XML as an enabling technology for sharing information. I'm not at odds with the authors of the aforementioned Scientific American article on this point. The article itself points out that XML alone is inadequate for the job. XML is great as a standard way of saying, "This next thing is a widget." But XML doesn't require that you describe what the widget does, how it works or that the widget itself conforms to a standard. 

RDF helps this situation a little. RDF is a complementary standard designed to describe the widget in terms of a subject, predicate and object. If I'm in danger of losing you at this point because you've forgotten your grammar lessons, don't worry, we're not going there. If you're curious about the details, visit www.w3.org/RDF. 

Otherwise, think of it this way: When XML points to a series of digits, RDF makes it easier for your computer to figure out that this is the phone number it should dial. Unfortunately, if one Web site uses the tag "cell" and another uses "mobile," RDF doesn't do anything to help your computer understand that they're basically the same. 

That's where ontologies come in. Ontologies are Web pages that contain a mystical unifying force that gives differing labels common meaning, even when crossing international language boundaries. We empower these pages by holding hands during each vernal equinox while we sing "Kumbaya." 

If astute readers detect a note of sarcasm, I plead guilty. It's not that the concept is unsound. It's just that there's this thing about human nature: Give folks a loose standard and the first thing many of them do is exploit its weaknesses for their personal gain. And as standards go, used-car salesmen register higher on the strict-o- meter than XML and RDF, never mind this nebulous concept of ontological Web pages. 

People thought nothing of repeating a word hundreds of times in HTML metatags to trick dumb search engines into giving them more prominence in the early days of the Web. The Semantic Web doesn't change the fact that these same people manage sites that will compete for your dollars. 

Don't get me wrong. I love the idea. For one thing, the Semantic Web describes a world where computer automation isn't an oxymoron. Given the processing power at our disposal, it's criminal that today's users should know what disk drives are and are required to recognize an executable e-mail attachment in order to avoid a Trojan horse or virus. 

The other thing I like about the Semantic Web is that if you step back and get a holistic perspective of the whole thing, you'll see that it's really a distributed, text-based object-oriented relational database management system (OORDBMS). No kidding. Have a look at the specs, and you'll find objects, properties, methods, inheritance, relational hierarchies - the whole enchilada. And I'm a big fan of OORDBMS technology. All I want to know is how you motivate millions of database administrators to play by the rules, loosely defined or otherwise. 

Nicholas Petreley is a computer consultant and author in Hayward, Calif. He can be reached at nicholas at petreley.com.