W3C home > Mailing lists > Public > www-tag@w3.org > January 2003

Re: Clarifying what a URL identifies (Four Uses of a URL)

From: Sandro Hawke <sandro@w3.org>
Date: Fri, 24 Jan 2003 08:10:20 -0500
Message-Id: <200301241310.h0ODAKF27952@wadimousa.hawke.org>
To: "Larry Masinter" <LMM@acm.org>
cc: www-tag@w3.org

> > >    indicates(context) (URI)  -> Concept
> > 
> > I'm a little fuzzy on what you mean by context.   Would it be fair
> > to simply have several different kinds of indicator functions and drop
> > the notion of context?   
> I was basing the idea of 'context' on the text that started
> this:
> http://www.w3.org/2002/11/dbooth-names/dbooth-names_clean.htm#DifferentC
> ontext
> but with the idea that there were many different contexts,
> not just four.

Right, thanks.   That was like 100 messages ago and I'd forgotten it.

Also, I think RDF is just one context, or perhaps the predicate
position is one context and the subject and object positions is one
context.  Anything like "a different context for every triple" breaks
RDF, I think.

> I'll give some examples of what I mean by 'context':
>   "XML namespace name":  The context of appearing in an xmlns
>   attribute.
> This is an important context in the W3C. XML uses URIs to
> indicate a 'namespace'. In this formalism, you don't have
> to imagine that a 'namespace' is a 'resource'.
> So
>     indicate(xml-namespace-name) 
> is a function that maps from URIs to namespaces.
>     indicate(xml-namespace-name)("http://www.w3.org/2000/svg")
>        indicates the abstraction "the namespace for SVG".
> On the other hand, XML encryption uses URIs to
> name encryption algorithms. So perhaps 'XML encryption
> algorithm' is another context. And
>     indicate(xml-encryption-algorithm)
> is a function that maps from URIs to encryption algorithms.
> indicate(xml-encryption-algorithm)("http://www.w3.org/2001/04/xmlenc#rsa
> -oaep-mgf1p")
> is a particular algorithm. An algorithm need not be
> a 'resource'.
> You asked some questions about what a 'resource' might be,
> and of course, it seems circular to answer that a resource
> is whatever a URI identifies. (Not 'indicates' or 'denotes',
> but 'identifies' in the sense above.) This really would
> be circular, except that it is (at least supposed to be)
> part of the definition of the scheme.
> That is, each URI scheme, when defined, should define
> what a URI using the scheme identifies. The range of
> identifies("http:" + string) is different from the range
> of identifies("mailto:" + string), but are defined
> (for better or worse) in their corresponding scheme
> definitions.
> If you want a more precise definition of what it is
> exactly that a HTTP URI identifies than you already have,
> you should then update RFC 2616, not RFC 2396.

Absolutely.  I was trying to ask what you thought an HTTP resource
might be.  (And is that the same as your answer to httpRange-14?)  Not
for the TAG's debate, but because you seemed to be expressing a
viewpoint I've heard mentioned but haven't understood.

One interesting point to consider is the difference between http:,
https:, ftp:, gopher:, ... resources.  It seems to me (and many others
here, I think), that they may be the same, or have a common superclass
at least.

(I'm happy with the use of the word "resource" here, since you've made
it clear that it's meaningfully constrained.)

  -- sandro
Received on Friday, 24 January 2003 08:12:38 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:55:57 UTC