RE: Do resources have representations?

> Then how can it be that two URIs that denote the *same* resource can
> give me *different* representations when I put them into my browser?

That could also happen if you put the same URI into your browser and changed
something else (say which languages you said you preferred).

In general this is also true of any identifier. And I do love analogies...

If you say "Jon Hanna" to one person he will say "ah, smart guy", if you say
"Jon Hanna" to another he will say "what an idiot" (hopefully the former
will occur more often than the latter).
A third person will think you mean John Hannah the actor, or a different Jon
Hanna, or not know of me at all. URIs prevent the first two mistakes (by
being globally unique), but not the third (which is analogous to a 404 or a
DNS-lookup failure).

Now, if you use "Talliesin" with the same group of people, some will
identify "Talliesin" as one and the same as "Jon Hanna", some will not
recognise it at all, or recognise it even if didn't know of me by the other
name.

Interestingly some people have (and maybe some still do) known me by both
names, but not known that the same person was referred to. Also since I use
those names in different contexts someone who knows me by both is likely to
give different information about me depending on which name is used. They
are applying semantics that are not necessarily innate to the identifier.

This relates back to URIs and the web in different ways depending on whether
we treat it as a collection of applications or a single application.

In the first case different servers know different URIs and different things
about the resource they identify. As such they inevitably give different
representations.

In the second case knowledge of the two URIs does not necessarily entail
knowledge of the fact that they identify the same resource. Even if this
knowledge is available the information given (i.e. the composition of the
representation) may be influenced by the identifier used, but this is a
property of the application's (the web's) action upon the identifier, not of
the identifier or the resource.

Received on Monday, 28 July 2003 11:33:15 UTC