Re: URI: Name or Network Location?

On Jan 21, 2004, at 17:15, ext Sandro Hawke wrote:

>
> This may be obvious, but "location" (as in URL) has nothing to do with
> physical location.  "http://www.w3.org/" names a location in
> information space, a location on the web.  That "location" isn't in
> physical space, isn't on any particular machine, etc.  (For people who
> don't know, service at the address is provided by a set of machines on
> three continents.)  The word/concept of location here is important
> because our human notions of persistence with respect to location
> match what works well on the web.  Our human notion of naming matches
> less well, IMHO.
>
>         -- sandro

Quite true. But insofar as implementations are concerned, the bits are
stored somewhere.

Using the same notion of "location" for both an abstract location of
some resource in an information space and the physical location
of the resource (or representation) in an 
implementation/server/repository
risks further confusion.

Above the line of implementational opacity, we can speak of names, which
denote resources, and which may be resolved to representations of those
resources.

Below the line of implementational opacity, within the context of 
particular
systems/solutions, we can speak of locations where representations 
reside
(even if those locations are "virtual" locations such that the 
representations
are always generated on-the-fly).

Both metaphors/perspectives are useful, but IMO on different sides of
that line between abstract web architecture and actual implementations.

Patrick

--

Patrick Stickler
Nokia, Finland
patrick.stickler@nokia.com

Received on Thursday, 22 January 2004 04:38:48 UTC