Lists Home |
Date Index |
> > Does that make sence?
> IMHO, no --dal
I agree... My grammar, spelling and logical structure are really bad. Sorry
about that! Let me clarify.
In Response To:
> > > I'm still trying to figure out /why/ you would even want to (use the
> > > HTTP scheme for things which you don't intend to be dereferenced).
> > > Whether it's harmful or not - has anyone actually given a reason for
> > > wanting to do this?
What I said:
> > You shouldn't (opinion alert). A URN would probably work better? On the
> > Other hand the "name-like authority" for URLs is well established
> > the authority for URNs) and it is a de-facto standard as well as an
> > one.
What I mean:
The "http:" scheme identifies the address of a resource on the net
(obviously). By analogy, the "postal address" scheme (not URI, but still a
standard resource location identifier) identifies the physical location of a
place. For example:
2) 1600 Pennsylvania Avenue NW, Washington, DC 20500
3) George Bush
"1" identifies the web site for the George Bush; "2" identifies the physical
location for George Bush's home; "3" identifies the name of George Bush.
Everything understood? (I'm being simple so that I'm clear.) Another George
Bush may live at the postal address:
3) 256 Dream Land NT, Fantasy Island 54321
4) George Bush
Here, "3" identifies the physical location of this George Bush and "4"
identifies his name.
How do we tell them apart? It is common practice to distinguish a person
via their address if their names are the same (I know from experience that
medical clinics and insurance agencies do this all the time). Thus we can
identify the first George Bush via his postal address "2", while the second
can be identified by his postal address "3".
In both cases, it made more sense to identify their customers by their
address rather than their common name (since addresses are unique by the
government and names are not). Since URLs are guaranteed to be unique by
the "government of the internet," and there is no body regulating and
keeping track of URNs right now (that I know of), then it makes sense to
prefer them over some URN to identify resources that may be retrievable via
Note "may be retrievable" and not "should be retrievable." The URL could be
reliably used to confirm the identity of the resource (just as the postal
address is used to confirm George Bush's identity) even though it isn't
intended to be dereferenced, such as when the identity of a particular
resource changes the behavior of another resource.
"...such as when the identity of a particular resource changes the behavior
of another resource." Here's an example: when XML 1.1 comes out, it might
get a namespace for pre-defined attributes (such as xml:lang, xml:base,
xmlns and others), which is different from XML 1.0. Then an XML processor
can go into "XML 1.0 mode" or "XML 1.1 mode," depending on the identity of
the declared namespace. In this case, the document describing XML 1.0 or
the XML 1.1 scheme won't be retrieved via the web; it's simply not needed
for the processor to do its job.
Does that make more sense?
"my mind is slipping away...
...day by glorious day"