RE: New AFTF draft.

I think the main point is that a URI doesn't imply "location", it is
merely an abstraction of something that has identity (I think this is
consistent with [1]). The notion of "server" is not in any way integral
to this, as I said in my previous mail [2], a server (or servers) may
show up in particular resolution mechanism but that is orthogonal to the
identifier itself. There is no dependency on client/server or otherwise.

Hope this makes sense,



>Hmm.  I don't think we can have it both ways.  If resources 
>don't travel, 
>then it's incoherent to say "this resource lives at server X". 
> In other 
>words, if something has a location, then it's necessarily 
>reasonable to 
>say that it has a location that moves, at least in a system like this. 
>Keep in mind that with my earlier example, there was on 
>picture of my on 
>any server.  The only state representing the picture at all was in the 
>message.  So, I think we can have it two ways:
>1) Resources don't travel because they are never localized.  
>They are just 
>abstractions with no notion of location or proximity.  In this case, 
>Henrik's proposal that resources don't travel makes sense.
>2) Resources do or often do have a location.  In this case, I 
>think it's 
>as coherent to say the location is in a message as on a 
>server.   I hope 
>we aren't tieing the notion of URI and resource to client/server?

Received on Wednesday, 11 September 2002 14:02:13 UTC