RE: fragment identifiers and media types (was RE:XPointerconsidered incomprehensible)

> From: Dan Connolly [] 
> On Wed, 2006-09-06 at 13:59 -0400, Booth, David (HP Software - Boston)
> wrote:
> > . . .
> > Protocol messages may be used in the process of determining
> > the meaning, but there is a difference between the meaning
> > being determined by the content of the retrieved message
> > versus the mere fact of retrieval. In determining the meaning
> > of , if a GET on
> > returns a 200 OK and an HTML document, the mere
> > fact of retrieval indicates that
> > identifies a location within an HTML document. It therefore
> > cannot, for example, identify a person or a dog.
> So don't publish an HTML document there if you want to use it
> to identify a Dog.

That was my point!  Hash URIs are more restrictive than slash URIs with
303-redirects because the meaning of a hash URI depends on the media
type.  I.e., they are interdependent.  Slash URIs with 303-redirects do
not have this limitation: you can serve any media type *independent* of
the kind of resource that you wish to identify.  

>From a design perspective, this difference is significant.  If you are
minting a new URI to identify something other than an "information
resource"[], it doesn't make any logical sense to tie the meaning of
that URI to the current media type of your documentation.  At some point
in the future, you may wish to serve your documentation using some new
or different media types.  (As the WebArch says, new media types are "a
means by which the Web can grow"[2].)  And as this thread has
illustrated, the question of whether the meaning would remain
"sufficiently consistent"[2] across media types is not at all obvious.

[2] WebArch section 3.2.2 on Fragment identifiers and content

David Booth

Received on Wednesday, 6 September 2006 19:49:00 UTC