Hi Phil,
> The DC in HTML recommendation [1] does not discuss the
> meaning of link metadata with rev attributes, all the
> examples are for rel attributes. The implied meaning is that
> the current document is the object of a statement whose
> subject is given by the link's href attribute.
>
> In a document whose URI is "http://example.org/current.html", this
> link:
>
> <link rev="DC.Source" href="http://example.org/other.html" />
>
> Gives the triple:
>
> <http://example.org/other.html>
<http://purl.org/dc/elements/1.1/source>
> <http://example.org/current.html> .
>
> Okay so far?
Yes, I think so.
> If this is the case, then the current document could make statements
about other resources that are not governed by the owners of the
> respective resources,
Yes, but that's one of the fundamental principles of resource
description on the Web: anyone can say anything about anything. I can
use Dublin Core to say I am the creator of "Lord of the Rings" or "Pet
Sounds" (in HTML or in some other format). Whether you trust/believe
that statement is up to you. That's the nature of the Web. And as you
suggest below, making that assessment leads you into making evaluations
based on the context of the statement, who made the statement and when
and so on.
But I don't think we can stop people making the statements - and I don't
think we should try!
> e.g.
> <link rev="DC.Rights" href="http://www.gutenberg.org/etext/5000" />
>
> This says the current document declares the rights to the notebooks of
Leonardo da Vinci, which is not true. The publisher of the current
document
> does not have the right to declare rights over another resource. The
publisher of the other (Leonardo) resource cannot control the metadata
statements
> made by the first (except legally).
First, I think the dc:rights property is maybe a bit more
general/"weaker" than you are suggesting
"Information about rights held in and over the resource."
Also, in your example, the current document (with the link in it) might
well make a true statement, even if it's not Project Gutenberg who makes
it!
e.g. I could say in my document "These representations of the notebooks
of Leonardo da Vinci are owned by Project Gutenberg and can be used only
in accordance with the conditions specified by the Project in document
xyz". We're not all out to deceive, you know. ;-)
Equally, as you say, I could indeed lie through my teeth!
> Does this lead us into the territory of reified statements? Document X
_says_ it declares the rights of document Y?
There are various approaches to the trust/provenance/context issues and
reification is one of them (but others here are better placed to comment
on that work than me!)
> Is this the intention of the DC in HTML recommendation, or should rev
links be prohibited for this reason?
Basically, no, I don't think they should be prohibited, and I don't
think we could do so anyway ;-)
But yes, I agree that means of establishing trust/authority etc are
important - after all we've all had to learn to be careful with
human-readable info on the Web and recognise that that ever so
polite-sounding Nigerian gentleman probably isn't going to give us a
share of his uncle's oil fortune, so ultimately the same applies to the
machine-processable stuff.
Cheers
Pete
|