RE: Non-Text Literals

>At 2002-09-01 11:16, Bill de hÓra wrote:
>>Ok. Self contained in RDF is handing someone an RDF graph. Sending
>>resource representations around with the graph is extra.
>
>Think of the image as a literal rather than a "resource representation",
>just like those bits of text that pepper most RDF.

An image we've encoded into a jpeg should be a resource, especially as it
is likely to contain metadata of its own, and people are likely to want
to say lots more about the image.  i.e. you say it's a photo of you, I
might want to say that it's a particular fine example of the
photographers art in my lump of RDF.

>True, but that's really the same as any literal. If I really need to
make
>statements about some image, I can always do something like this
>(assuming I can specify MIME-type and encoding):

Certainly you can say things about it in your document, but I can't say
anything in it in mine, without repeating all the literal data (and using
some method to say it's unambigous.)

However packaging images and RDF descriptions in single packages is
something I've wanted to do to, I've found two approaches, the first
where the image is the important thing, I can embed the RDF inside the
image using Adobe's XMP method, RDFPic and Dan Brickley's online service
and Adobe's tools can all access this information.

Alternative I embed the RDF in SVG, I can include my images within an SVG
document, and also RDF, this gives me a well known and understood method
for having images and RDF within the one document.

An example of this  is - http://jibbering.com/2002/9/svg-rdf.svg (Where
I've taken the image and some RDF generated by
http://jibbering.com/svg/AnnotateImage.html ) Whilst I've used SVG I'd've
thought any method which allows you to package RDF and other
representations in a single document is preferable to literals, if
packaging is required.

Jim.

Received on Monday, 2 September 2002 06:21:14 UTC