Lists Home |
Date Index |
> I had held the hope that one of the things the Semantic Web would be
> good for would be to enable me to make useful machine-readable
> statements along the lines of "this resource is just a JPEG of my cute
> cat" and "this resource serves as a placeholder for the W3C in my KR
> system". But Pat Hayes et al tell me I'm all wrong. Oh well.
Hypertext links with arbitrary metadata in the markup of your choice can
achieve the first example:
<image src="http://www.tbray.org/ongoing/cute-cat.jpg" type="image/jpeg">
<animal nature="cute" type="feline">
My Cute Cat
Alternatively, embed machine readable properties in the image itself,
which is quite possible with JPEG and PNG images.
The second example is even easier, presumably:
... properties sufficiently describing W3C ...
Now if my KR system uses http://www.w3.org/Consortium or what have you,
the two systems can compare notes to see if they are talking about the
same thing, or at least sufficiently the same for the purpose of the
A URI is not a picture of your cat, but the stream of bytes that can be
retrieved using that URI with HTTP *is* :)
A URI is not the W3C, but I can use one to denote it if I state clearly
enough what I think I mean.
English says "fish", Japanese says "sakana" (if I recall correctly) yet
communication takes place by determining that these two words denote
sufficiently similar concepts, by comparison of the properties ascribed to
I think what Pat Hayes is demonstrating is that there is no uberword
http://fish.com that is the one true identifier for the concept of fish,
and that the intersection of REST with the SW -- while they may share the
same vocabulary (resources etc.) is schizophrenic.
YesLogic Prince prints XML!