[rfc-i] BCP 35 and BCP 115 and Auth48 and clerical errors

"Martin J. Dürst" duerst at it.aoyama.ac.jp
Wed Apr 15 01:38:30 PDT 2015

Hello Larry, others,

On 2015/04/15 13:43, Larry Masinter wrote:
> BCP 35 was RFC 2717 (procedure for registering schemes) while RFC 2718 was Informational (Guidelines for new URL Schemes).
> RFC 4395 obsoleted both 2717 and 2718, but got assigned BCP 115. Later (I forget the exact circumstances) RFC 4395 was made to also be BCP 35, but RFC 4395 still says it is BCP 115.
> I think the situation is confusing and there is no explanation for someone just looking.

There are BCP that consist of more than one RFC. The example I know is 
BCP 47 on Language Tagging (see e.g. 

> draft-ietf-appsawg-uri-scheme-reg will obsolete RFC 4395.
> Having two BCPs point to the SAME document is a problem.
> Pick ONE of BCP 35 and BCP 115 as the right BCP number for the new RFC. (I’d pick BCP 35).

This is already done if you look at 
http://www.rfc-editor.org/bcp-index.html, which has:

BCP35	Guidelines and Registration Procedures for New URI Schemes T. 
Hansen, T. Hardie, L. Masinter [ February 2006 ] (TXT = 31933 
bytes)(Obsoletes RFC2717, RFC2718) (Also RFC4395) (Status: BEST CURRENT 


BCP115	[BCP number 115 is retired. It was mistakenly assigned to RFC 
4395. RFC 4395 is BCP 35.] [ ] (Status: BEST CURRENT PRACTICE)

 > Replace the other one (BCP 115) with a note explaining the situation 
and refer people instead to the ‘right’ number (BCP 35).

http://www.rfc-editor.org/bcp/bcp115.txt currently returns a 404, this 
could indeed be improved.

> How could this have been prevented? Should the Obsoletes: RFC nnnn (if approved) include BCP numbers or STD numbers so reviewers can double check?

Yes, having obsoletes listing BCPs might help.

Regards,   Martin.

More information about the rfc-interest mailing list