--On Thursday, February 24, 2011 09:42 -0500 Andrew Sullivan
On Wed, Feb 23, 2011 at 01:59:23PM -0500, Dave CROCKER wrote:
I was impressed with just how steady that increase appears to
be over a reasonably extended period of time, as well as
its seeming to be around 70%, not 50%, now.
Pretty serious 'market' domination...
Sure. But, having started this thread, let me point out that I
don't think anyone has made a serious claim that xml2rfc is
unimportant to the community, that it should not work well, or
even that putting resources into making it more stable, more
maintainable, or better behaved is a bad idea. Those figures
are interesting but they would be relevant only if one of those
"don't do this at all" arguments were being made. They haven't
I raised two issues and two issues only:
(i) That the "tell the community work is going to be started by
issuing an RFP" model was suboptimal and maybe inappropriate. I
think the discussions and conclusions of the last two weeks
indicate that the community agrees with that position and that
the IAOC has recognized that agreement.
(ii) That perhaps the priorities about extensions in the draft
RFP were inappropriate. I want to stress "perhaps" because,
while I have a view on that subject, I think the correct
priorities can emerge only from community discussion (which
takes us back to (i) above). Interestingly, while the numbers
you and others cite can be used to argue that improving RFC
Editor Production Center Staff efficiency in dealing with
documents arriving in xml2rfc format, they can equally well be
used to argue that, if the community is producing that much text
in that format, it is at least equally (if not more) important
to improve the efficiency and capabilities of the tool (and
DTD/Schema) for those who are originating/writing the documents.
I submit that this is a consequence of the submission process.
The idnits tool is now an effective gatekeeper against people
who are preparing documents with something other than xml2rfc
and, maybe, the *roff gui that's available. As a WG chair of
a WG with a lot of participants who have older toolchains, I
field regular complaints about how difficult it is to get
things by idnits. My stock advice is to use xml2rfc, which is
always being updated exactly so that it conforms to the latest
physical layout rules.
I personally think this is sort of a shame, because it reminds
me of other (near) monocultures based on tools that are
considerably worse than the state of the art (Word is my
I agree, especially with regard to having the tools
over-constrain early drafts and sometimes thereby making it
harder to expose ideas and get work done. I think the community
should be pushing back, very hard, on submission tools that
enforce rules that have never been formally announced by the
IESG and/or approved by the community. But I think it is (or
should be) a separate issue from whether and how we improve the
behavior or robustness of xml2rfc.
Ietf mailing list