OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Names and locations (WAS: Why REST?)

[ Lists Home | Date Index | Thread Index ]

John Cowan wrote,
> But labeling me and my home page with the *same* URI leads to 
> contradictions.  For example, the year of creation of my home page 
> is 1998 or thereabouts, whereas *my* year of creation is 1958 or 
> thereabouts.

Only if you insist that these URIs (in this case URLs) must function 
as names.

But they don't, at least not necessarily. URLs can (and IMO should
primarily) function as descriptions predicated on a location in web-
space ... that's why we've traditionally, if controversially, made a
sharp distinction between Universal Resource *Locators* and Universal
Resource *Names*.

Treating http://www.ccil.org/~cowan as a name leads to contradiction,
as you say, because it implies that you and your home page are one
and the same thing. But treating it as a description by way of a
location is just fine: all that implies is that, in web terms, you and
your home page are in the same place right now.

Of course if you take the "location" and "space" metaphor too 
seriously you're likely to find this superimposition a bit spooky ... 
the medium sized dry goods we deal with most of the time tend to elbow 
each other out of the way. But web "space" is only a metaphor, and all 
metaphors break down when pushed beyond their domain of applicability.

Here's a few fairly obvious examples of things that this approach 
allows us to say easily and coherently. We can say that John Cowan
and his home page are at the same location, as above; we can say
that either or both have moved to http://www.cowan.org/; we can say
that the home page has gone, and a GET on that URL will return some 
other resource; we can say that the home page (that very same page) is
mirrored in multiple locations.

These all depend on an at least rough and ready distinction between
a resource and a potentially ambiguous means of referring to it, a 
distinction between identity and *identification*. I think that's an 
important distinction to hang on to, even tho' it's very hard to pin
down exactly what it is. I happily concede that names are vulnerable 
to ambiguity in some contexts, and that descriptions can identify 
uniquely in some contexts. I also happily concede that whether 
something is a name or is a description is often subtly context
dependent. But none of these considerations strike me as providing
sufficient grounds for collapsing the distinction.

It's tough dealing with ambiguity and context dependence, but the
problems can't be made to go away be defining them out of existence.
The Semantic Web, be it hubris or the next big thing, will be mired
in it ... but so will everything else, so we'd better just get used to
the idea and either tackle it head on or develop strategies for
routing around.



Miles Sabin                                     InterX
Internet Systems Architect                      27 Great West Road
+44 (0)20 8817 4030                             Middx, TW8 9AS, UK
msabin@interx.com                               http://www.interx.com/

This message is intended only for the use of the person(s) (the "intended recipient (s)") to whom it is addressed.
It may contain information which is privileged and confidential. 
If you are not the intended recipient, please contact the sender as soon as possible.
The views expressed in this communication may not necessarily be the views of InterX plc.
Any copyright in this message shall remain vested in InterX plc  and the intended recipient may only copy the same for internal business purposes or as otherwise stated in this message.

This e-mail has been scanned for all viruses by Star Internet.


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS