[rfc-i] DOIs and RFCs

Nico Williams nico at cryptonector.com
Tue Feb 11 15:50:50 PST 2014

On Tue, Feb 11, 2014 at 5:12 PM, Joe Touch <touch at isi.edu> wrote:
> There are two identifiers for each DOI object:
>         the DOI itself
>         the URL the DOI redirects to

Incidentally, there are several URIs for each RFC, one for each output
format and also there are two publishers (in the sense of URI
authorities), the RFC-Editor and the IETF.

> The former is useful only when the latter varies AND the former is updated
> accordingly. This may be the intent of DOIs, but the older they are the more
> likely they're broken more than they work in my experience.

I see.  I'm not sure what the relevance to us is that others' DOIs
break, unless that has led to DOIs not being widely used.  John claims
DOIs are widely used.

John argues that some organizations will have an easier time referring
to our documents by DOI than by URI (just as we have an easier time
referring to our own documents via series name and number).  Some
evidence of that would be useful, but I can believe it.

John also mentions bibtex implementations that can automatically
retrieve reference metadata via DOIs.  This does seem valuable to me.

> The cost seems small enough that I'm happy to ignore it.
> My concern is handing out more than one ID for a document. As the saying
> goes, "a person with one watch always knows what time it is; a person with
> two is never sure".

We already have multiple URIs for our documents, and even multiple
"IDs" (STDs, BCPs).  And URNs...

> I don't mind if we create DOIs for RFCs, but IMO we should only give out the
> rfc-editor URL, in which case there's little point in the overhead
> (organizational or financial) of using DOIs at all.

Or if we can give multiple URIs, then we should do that.


More information about the rfc-interest mailing list