Re: HTML Link Type Model

Roy T. Fielding (fielding@avron.ICS.UCI.EDU)
Thu, 18 May 95 00:35:11 EDT

Murray writes:

>> I still think the definitions I gave are more accurate -- both
>> REL and REV do specify specific source and destinations (assuming we
>> don't invent multi-ended links).
> I don't know about y'all, but we have multi-ended links already.

I wasn't referring to the affect of multi-ended links -- only the syntax.
The syntax can only define point-to-point links, one at a time,
but possibly with multiple relationships. It takes more than one
LINK or A to assign relationships to one-to-many, many-to-one, or
many-to-many links, because an href only points to one resource
(and the link is to the resource, not the set of possible results
returned from requesting that resource).

> This is currently accomplished through the use of a cgi-bin
> script to resolve a request for a named document of which there
> might be multiple occurences.
>> ... I'm saying that the WWW link model does not adequately define
>> the meaning of relationships based on noun phrases.
> We agree on that much. So, having abandoned the "WWW link model",
> can you tell me why SCO and the HTML WG shoul abandon REL/REV
> as SCO has implemented.
>> >
>> > Ah, I see... "REL" labels the link, while "ROLE"
>> > would label the *target* of the link. Correct?
>> Yep.
> But by abandoning REL/REV, you lose the expressive power
> provided by describing the relationship in both directions.

No, you merely make such relationships dependent on the availability
of inverse names. Is that desirable? Probably not. The useful question
is whether the gain from having a more direct, visible syntax outweighs
the loss of generality. The WWW model went for generality.

> Or we could abandon the use of verbs along with the "WWW link model"
> and just stick with nouns. Would that work for you?

No, as that would conflict with existing practice, and would be no
more useful than the existing set of affairs. There are advantages
to having both, and you can't stop people from using both, so you are
better off defining a model in which both can be used consistantly.

<LINK rel=Obsoletes href="rfc1321">
<LINK rel=Updates href="rfc822">

....Roy T. Fielding Department of ICS, University of California, Irvine USA