Re: sharing MIME types and other enumerations [was: notes on SLinkS RDF schema]

I've been busy fixing things...


At 2:16 AM -0600 11/10/99, Dan Connolly wrote:
>Eric Hellman wrote:
>> What, exactly, is the point of replacing the string text/html with the
>> resource http://www.isi.edu/in-notes/iana/assignments/media-types/text/html
>> , anyway?
>
>Hmm... good question. I suppose MIME types are just like country codes,
>ISBNs, etc. My intuition says that the way to manage these global
>enumerations is to ground them in URI space. But I can't put my finger
>on exactly why this is... here are some related thoughts:
>
>	"The Web works best when anything of value and identify is a first
>	class object.  If something does not have a URI, you can't refer to it,
>	and the power of the Web is the less for that."
>	-- TimBL, Dec 1996
>	http://www.w3.org/DesignIssues/Axioms



It's a bit bombastic to call this an axiom; I would say it's more of a
academic conjecture. Is it really that bad to call Drosophila melanogaster
a fly? I can refer to the Pope just fine without giving him a social
security number; he has both value and identity.

URI's are excellent at being unique; they're not so good for being
universally accepted or used as identifiers. MIME-Types are not a bad
example. The IANA URI's you cite are certainly unique, but they have a lot
of disadvantages versus the raw strings.

1. They are obscure. Before your comment I had never heard of them. (yes, I
admit to being a rare, ignorant dolt)

2. Relevant processors do not understand them. try putting a MIME- URI in a
MIME header and see how far you get.

3. They appear to be unsupported. By support I mean that no one appears to
have bothered to make the resources addressed by these URI's in the least
bit useful.

4. They do not have provenance. There is no statement of authority for
these URI's on the URI owner's web site. (that I could find)

5. They have no RDF content.


So, in order to get my RDF schema application to work, in a product that
needs to ship now, I have to stick to my local, stick-in-the-mud
enumerations, however much I would like to use something more universal.

This points up some basic infrastructure needed to make RDF Schemas work in
the real world.

1. Common, well-known resources need to be RDFied by appropriate
authorities. Examples:
	MIME-Types
	Countries (Country codes)
	languages (language codes)
	units of measure

2. A web page pointing at URI authorities (and RDF schemas, for that
matter) needs to be developed and publicized.

3. A system needs to be developed for schemata which get abandoned. This
might even take the form of a redirection site. Hmmmm. If a resource
addressed in an RDF document gets redirected during resolution, what
resource is pointed at?

4. A convention for indicating what to do with the resource at the end of
an RDF URI needs to be developed. This should take the form of a simple RDF
Schema itself.

Using URI's in stead of strings only gains you something if there's a
community that has gotten together to decide to use the same URI's for the
same concepts.

URI's are like sounds. It's only when people get together to assign meaning
to sounds that we get a language. (Actually, you need children getting
together to make a reallanguage, but that's another story.)

Eric

PS. I used most of your other suggestions to build a new version of the
S-Link-S schema
Eric Hellman
Openly Informatics, Inc.
http://www.openly.com/           21st Century Information Infrastructure

Received on Tuesday, 16 November 1999 02:45:22 UTC