On Mon, Oct 26, 2020, 6:06 PM John C Klensin <john-ietf@xxxxxxx> wrote:
Brian,
I look at the same information and come to a different
conclusion (quite independent of the question of whether a poll
at this point is a useful exercise)...
--On Tuesday, October 27, 2020 07:56 +1300 Brian E Carpenter
<brian.e.carpenter@xxxxxxxxx> wrote:
> As Julian Reschke observed on the rfc-interest list, since the
> new RFC format was implemented:
>
>> page numbers should not be used to refer to parts of the
>> RFC, because page breaks vary with output formats
Cross references (and other references) to page numbers have
been discouraged since at least RFC 1543 (in 1993) and banned
soon after that, to the point that the RFC Editor from the last
half of the 1990s would simply remove such references, replace
them with references to section numbers, and complain when
sections got too long for convenient referencing. Nothing about
restricting references to page numbers is new with the new
format.
> So I can only see confusion if people use page numbers for
> any purpose whatever. So it doesn't matter if people want
> page numbers; they're now useless. So I won't be answering
> a poll, and I don't think the results are interesting.
However, getting from "references to information by page numbers
have always been a bad idea, prohibited for a quarter-century,
and even more obviously a bad idea as we move to multiple
formats" to "any purpose whatsoever" is a big jump. At least
some of us have tools and macros floating around that are
dependent on pagination and, as an exception to the
"referencing" rules, it is still not clear (at least to me) how
to build and format a document index using anything else (at
least without a lot of effort). There are even simple and
obvious reasons: If one is going to print an RFC from the text
form (or render it into printable form), something is going to
do the pagination and being able to easily estimate the page
count may affect how printing is to be set up.
FWIW, the questions of "should documents be paginated" and
"should the pages be numbered" are also separate ones.
Moreover, the argument that pagination (and page numbers) are
obsolete and useless for RFCs would be much stronger if the PDF
form for new RFCs were not paginated (or at least not numbered)
... but it is both paginated and numbered. And, if the issue
is having things lay out well on many different types of
devices, eliminating pagination (and headers and footers) to
facilitate that is bogus: it would be equally or more reasonable
to eliminate (or rethink) line breaks in running text, etc. If
one really wants things optimally formatted for a variety of
different devices, then the right thing to do is to start from
the HTML form and a well-designed style sheet or to go back all
the way to the XML, not to try fussing with the ASCII text form.
Conclusion: The main arguments that have been given for
eliminating pagination, headers and footers, and page numbering
--at least those based on different devices and references-- are
mostly bogus.
So, from my point of view as a fan of pagination in the ASCII
form of RFCs, one who has never willingly referenced part of an
RFC by page number, this seems to come down to something else
entirely: if there is a goal to eliminate (or strongly
discourage) the use of ASCII format RFCs in favor of the HTML or
PDF forms (or building directly on the XML), then "no page
numbers" and "no pagination" are among the first few cuts of a
death by 1000 of them. If not, this has all of the hallmarks
of a gratuitous change to a format that has been useful to many
people for a very long time.
john
p.s. Just in case I'm the anonymous person John Levine's note
referred to, I didn't just "not participate" in the discussion.
It was made extremely clear that my input was not welcome, so
clear that, after a discussion or two with Heather that I should
spend my time in other ways. If there were a significant number
of others like me (and I have no way to tell, or even to guess)
then claims that the no-pagination form represents community
consensus lie somewhere on the scale between "dubious" and
"bogus".
Disagree. It is far past bogus. How about "somewhere between 'false' and 'disingenuous'"?
