Tim Berners-Lee
Date: @@@@, last change: $Date: 2009/08/27 21:38:09 $
Status: personal view only. Editing status: first draft.

Up to Design Issues


My Top Ten Terms

"A little semantics goes a long way", says Jim Hendler. He uses the phrase in lots of ways. I'm going to use it to say that the sense that a few ontological terms are so common that they will pull together all kinds of data ion all kinds of places. ___@@ at Mitre calls this "loose coupling: the idea that you would connect things through these few terms, and leave other connections to future projects. He talks about "Who, where, when?"

These are my ten terms for todays selected on the basis that they are important common dimensions, and if more data was connected using more of them then the data in the world would be a whole lot more connected. They are chosen from a pragmatic standpoint. There are probably

People

Let's start with "Who?" . I've picked two terms from the Friend of a Friend ontology because it is an ontology designed and used specifically in a bottom-up way to connect many people through nets of "knows" links between acquaintances.

foaf:mbox

The email address is the practical way most people on the net are identified. Web sites use email callback to establish that someone really does have the given email address. Yes, it isn't perfect, as

foaf:name

"When"? There are large numbers of quite complete ontologies of time, which use different models to discuss the complexities of the different scales, and of things like

cal:dtstart

cal:dtend

"Where" is a huge one.

geo:lat

geo:long

Documents are of course the lifeblood of the Web and the Semantic Web. The first call for structured data, which gave rise to RDF, was for data about documents. The Dublin Core group sat down to pick their top ten properties for docuemnts. Of the ones they produced, dc:title is the one I have found the most call for. I take the html <title> tag to have the same meaning.

dc:title

foaf:maker

In a fractal world, the total cost of ontologies for any project is reduced bny the fact that ontologies already exist in various commiunities


Up to Design Issues

Tim BL