Re: Request to register "identifier" relation type

[Bjartur Thorlacius <[e-mail]>]

The short version (as per Ed's separate email) is that "canonical" exists
to resolve URI aliases: 2 or more URIs for the same resource (i.e.,
duplicate content).  [...]

In contrast, "identifier" is about establishing relationships between
*different* resources.  This includes providing machine-readable linkage
between:

https://doi.org/10.1371/journal.pone.0171057

and

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0167475

but also between the DOI and the subsequents PDFs, PPTs, PNGs, XML, etc.:


If you discovered one of the three above URIs out of context, it would be
nice if they expressed membership in "10.1371/journal.pone.0171057".  That
wouldn't preclude additional boundary expressions, but the XML citations
and PNG of an article image clearly can't use "canonical".

If the content is different parts of a paper, then the subpages should
<link rel=up hef="/to/the/paper" >, should they not?

The DOI is the canonical IRI. There may be a couple of alternate
representations of the whole paper (HTML, PDF,  and paper). Then there are
parts with their own IRIs (images, tables, even chapters). The parts should
link up to the DOI (explicitly using rel=up).

To me, rel=identifier sounds like a hypernym of alternate and up
specifically intended for pointing to the nearest DOI. If it's only
intended for DOIs, why not just call it rel=doi? In non-DOI instances the
distinction between alternate and up seems important.

To be fair, RSS and Atom go so far as using alternate where index would be
more appropriate. So specifying the desired semantics explicitly, under a
new rel, and then moving to existing ones seems like a more
forwards-compatible strategy.

Regards,
Bjartur Thorlacius