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

Andrew Sullivan ajs at anvilwalrusden.com
Fri Jun 22 12:51:24 PDT 2012

On Fri, Jun 22, 2012 at 07:29:37PM -0000, John Levine wrote:
> Of course, but despite the availablility of well written C and Fortran
> standards, there's still plenty of traffic in comp.lang.c and
> comp.lang.fortran.

Ok, but (just to pick on the thing that got in my way most recently)
the XML xml2rfc accepts has an inflexible and poorly-considered way of
specifying any bibliographic references that aren't RFC-like.  Right
now, we can work around this because we know the final document isn't
going to be the XML we submit, but some other document that is
produced by additional tools.  This would be like C working perfectly
well as long as you ran it first through a special black box.  (So,
you know, like Java ;-) If the source file has to be the canonical
file, the bar for the tools goes way up: we can't shrug our shoulders
about corner cases and say, "Well fix that up by hand," because there
is no "by hand" step.  That's all I mean.  This is not a point about
xml2rfc, as Paul noted upthread, but a point about any case where we
make input-file-to-tool the canonical version of the text.



Andrew Sullivan
ajs at crankycanuck.ca

More information about the rfc-interest mailing list