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

Joe Touch touch at isi.edu
Sat Jul 7 08:57:24 PDT 2012



On 7/7/2012 8:19 AM, Paul Hoffman wrote:
...
> However, no one here is proposing that the canonical format for RFCs
> use just "the basics of HTML". All of the proposals so far have included
> at least two significant features that only regular HTML users and
> experts would consider "basic": attributes and the mandatory closing of
> tags. Probably more than 90% of HTML users in the IETF do not normally
> use attributes, and would have to be taught the rules for them,
> particularly quoting rules. Probably only half of HTML users in the IETF
> know that you really should close all your tags.

So we should be doing things for the reason that we "really should".

How about a reason tied to a function *required* for RFCs?

AFAICT, there are only a few features we really need:

	A- to mark and refer to locations within a doc
	(i.e., internal links)

	B- to indicate headers and lists

	C- to identify a very few document components for
	extraction (e.g., title, authors, date, RFC number,
	and abstract)

Attributes in a doc except for (C) are unnecessary (and IMO ought to be 
stripped out).

Closing tag requirements except for (B) are unnecessary (and IMO ought 
to be optional - as they continue to be in HTML).

Making our requirements exceed that of HTML serves no useful purpose 
except as an exercise for programmers - which this process ought to 
avoid, rather than encourage.

Joe



More information about the rfc-interest mailing list