[rfc-i] DOIs

"Martin J. Dürst" duerst at it.aoyama.ac.jp
Tue Jan 26 19:18:10 PST 2010

I'm working in an academic context in Japan. I'd say that most of our 
faculty and even more of our students never heard about DOI. When 
searching, it's first everyday search engines, then maybe Google 
scholar, then the library-specific stuff.

For me personally, as mentioned, DOIs are associated with bean counting, 
and we don't need bean counting for RFCs. I wouldn't mind if somebody 
worked on getting DOIs for RFCs for those who like DOIs, but I'd oppose 
to declaring them as "canonical" for RFCs.

I once attended a talk by Bob Kahn. Very interesting, but there was one 
thing where I couldn't disagree more. He presented the way content 
should be dealt with as if bean counting in doi: style were the correct 
thing, and open accessibility as the Web brought it was mostly a 
mistake. I'm not sure I remember correctly, but I think he pointed to 
DOI as the way to fix things. I don't mind some people or organizations 
using the bean counting style, but

Somebody mentioned urn: doesn't work on Safari. I guess doi: won't 
either. Looking at 
https://datatracker.ietf.org/idtracker/draft-paskin-doi-uri/, it seems 
an attempt to get a doi: URI scheme failed, but I might be wrong, and I 
don't know why. doi: is mentioned in a few places, e.g. 

On 2010/01/27 10:46, Julian Reschke wrote:
> Henning Schulzrinne wrote:
>> I don't know where you work, and this may well be more of a US/North
>> American thing than a European one (although Springer, being more of a
>> European publisher, uses them for their journals, too), but they are
>> pretty standard currency in academic research, which is presumably the
>> area where long-term-useful citations matter.
>> ...
> I'm working on Internet stuff. Surprise.
> The question that triggered my request was raised by the W3C, btw.

For W3C purposes, what I would currently recommend is 
http://tools.ietf.org/html/rfc[0-9]+. This is really great in that it 
shows a faithful representation of the official ASCII version while also 
providing a bit of metadata and linking wherever possible (inside the 
document, to other documents, and from other documents to specific 
pieces of the document). It also prints very nicely with a lot of 
software (even modern stuff that no longer knows about form feeds).

There are still however some problems with this:
- It's on the 'tools' site, which is getting more and more convenient 
but doesn't have that much of an official standing.
- It has /html/ in the path, which is as wrong as .txt in the extension.
- It's created by a mechanical process that not always gets things 100% 

So in my view, the ideal would be:
- http://www.ietf.org/rfc/rfc[0-9]+ and/or 
http://www.rfc-editor/rfc/rfc[0-9]+ serve the format currently served at 
tools.ietf.org together with the .txt format with content negotiation.
- This (these) are declared canonical locations

Regards,   Martin.

#-# Martin J. Dürst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst at it.aoyama.ac.jp

More information about the rfc-interest mailing list