[rfc-i] Does the canonical RFC format need to be "readable" by developers and others?

Paul Hoffman paul.hoffman at vpnc.org
Fri Jun 22 08:51:23 PDT 2012

On Jun 22, 2012, at 8:43 AM, Andrew Sullivan wrote:

> On Fri, Jun 22, 2012 at 05:14:48PM +0200, Julian Reschke wrote:
>> On 2012-06-22 11:53, Iljitsch van Beijnum wrote:
>>> The problem with XML2RFC is that it makes perfect sense once you're an expert, but then, you first need to be an expert.
>> There's a mailing list where you can ask for advice.
> If we are going to pick the Official Format of Some Tool as the
> canonical version of a document, the documentation for that tool
> cannot be "ask on the mailing list for help".

Correct. Julian wasn't saying that that was the long-term answer. He was saying that undocumented answers are available now for the current format.

>  I think I have fewer
> complaints than Iljitsch about xml2rfc, but I do not believe that the
> documentation model of Linux systems circa 1995 is an acceptable one
> for the mainline standard tool, if it's to become such.

Fully agree. The same would be true regardless of the input format: there has to be complete documentation, including hand-holding for the 25% of RFC submitters in the lowest quartile of understanding of how to do this. That's part of the RFC Editor's job.

> I still refuse to state an opionion of whether some official format
> convenient to one tool ought to be canonical.  But if it is to be,
> then I think we must insist that the format has to be able to handle
> corner cases that, perhaps, it does not at present.  Otherwise, the
> required format won't really be canonical -- some other form will be.


--Paul Hoffman

More information about the rfc-interest mailing list