Re: URIs, used in RDF, that do not have associated documentation

Hi Danny,

2012/3/28 Danny Ayers <>:
> 2012/3/28 トーレ エリクソン <>:
>> I did mean what I said. Classifying URIs as content-oriented vs.
>> description-oriented is in my opinion the root problem. What I want to
>> say is:
>> * All HTTP URIs are description-oriented, even if they return a 200 *
> I'm inclined to agree with you about the root problem, but I don't
> think choosing one side over the other helps solve it.
> The definition of an IR, that it can be totally represented in a
> message is at odds with WebArch. A resources has a set of
> representations, there isn't any single representation given special
> status as "The One". It may exist, and this may be stated
> somewhere/somehow. (There is also the possibility of media types we
> haven't yet invented, cf. the hi-fidolity "thing/dog").
> From this I'd suggest it follows that we can't assume that a 200 means
> we're getting the canonical message. It could perhaps be argued that
> it's all very conceptual, that the set of representations considered
> as a whole contains the message. But I don't think this gets us any
> further in practice.
> Talking about description-oriented is an improvement in the sense that
> it gets rid of the need for "The One". But I reckon on the way it
> loses the notion of a representation. It seems easier to approach it
> from the other direction and say that a description is an
> approximation to content. Take a photograph of a traffic cop, It isn't
> a traffic cop, it's a graphic description of one. But you set that
> photo at the side of the road and cars will slow down. The description
> isn't the thing, but both can have the same effect.

We agree completely. An img/jpeg octet stream can be a used as a
representation of the traffic cop, as well as the photo.
Representations are reusable.

> So I'd suggest a 200 would be legitimate for a description (in
> addition to what we normally refer to as content), because
> descriptions can be treated as representations.

Sure, that was my initial way of thinking of the problem, but other
people vehemently opposed that an img/jpeg octet stream could be a
direct representation of a human. Switching to pure description-based
semantics neutralizes this opposition. The end result is the same


Received on Wednesday, 28 March 2012 21:57:59 UTC