Re: [art] New RFCs text formatting

["Scott O. Bradner" <[e-mail]>]

I agree with what John has to say but also with the “bumper sticker” way that Keith
expressed it - changes should be made when to very well established and widely used systems 
its known that no one uses the old version not just because someone thinks there is a better way

for what its worth - I have used the old format for decades and still do (or want to)
i.e. pages & page numbers

Scott


> On Dec 1, 2019, at 3:23 PM, John C Klensin <[e-mail]> wrote:
> 
> I would say "+1" too, except that, from experience, the implicit
> characterization of paginated (and header and footer-equipped)
> plain text as being only about printing (especially, as has come
> up in other notes, printing on ancient, perforated-paper, line
> printers) misses the point that there are at least three reasons
> that I can think of to prefer that format (other than
> idiosyncratic personal taste, but I don't think the IETF gets to
> say that even the latter is illegitimate).  Those are:
> 
> (1) One really does intend to print the document and has a
> printer, or printer drivers, that make handling of documents
> without page breaks difficult.  As Paul Hoffman just pointed out
> (and I pointed out last week), RFC 7994 requires pagination (but
> just page breaks, not headers or footers) for the RFC output.
> If RFCs are being produced and put into the archives / available
> from the RFC Editor's server) that don't have those page breaks
> in the plain ASCII form (and we seem to have at least one
> example) that is a bug and it is, at least IMO, reasonable to
> ask how soon it will be fixed (and maybe why it happened).
> 
> (2) One has discovered that, regardless of the format in which
> one prefers to read RFCs, it is easier to work with them in
> plain text format using public tools like "grep", some
> highlight-copy-and-paste operations, and so on.  Most, perhaps
> all, of those uses require that plain text files be available;
> few, if any, require pagination, much less headers, footers, or
> page numbers [1].
> 
> (3) One has private tools (or obscure public ones that there has
> been no evidence that anyone other than their authors know
> about) that are actually dependent on the presence of headers
> and/or footers and/or usable page numbers.
> 
> So, I think that Keith should have said "only if there's rough
> consensus that
> none of the three cases above exist anymore.  If they do, there
> is rough consensus that the IETF is willing to tell users whose
> work falls into one of those categories that we are just not
> interested in their problems and that, if the result is that
> they reduce their participation or willingness to accept RFCs
> for other purposes, we just don't care."
> 
> best,
>   john
> 
> p.s. Some of the comments about A4 paper make questionable
> assumptions about how line and page lengths for RFCs were
> chosen.  It may well be that we (and I do mean "we" because I
> was involved in the discussion although I certainly was not the
> decision-maker) got it wrong, but the assumption that is was not
> considered is simply incorrect.
> 
> [1] As I have discussed elsewhere and will not repeat unless
> asked, the claim that eliminating the ability to reference into
> an RFC by page number and therefore encourage shorter sections
> has actually been disproven by years of experience.  The only
> things that encourage or forces shorter sections are author
> education and the RFC Editor insisting that they will not
> publish a document with too-long sections without a very
> specific justification (at least unless the Stream Manager
> insists and then probably with the appropriate note).
> 
> 
> --On Sunday, December 1, 2019 12:55 -0500 "Scott O. Bradner"
> <[e-mail]> wrote:
> 
>> +1
>> 
>>> On Dec 1, 2019, at 12:52 PM, Keith Moore
>>> <[e-mail]> wrote:
>>> 
>>> 
>>> 
>>> Sent from my iPhone
>>> 
>>>> On Dec 1, 2019, at 12:28 PM, Nick Hilliard <[e-mail]>
>>>> wrote:
>>>> 
>>>> I'd be happy for pagination to be the default if there were
>>>> rough consensus that the primary medium for reading RFCs was
>>>> printed paper.
>>> 
>>> Seems like this is backwards.  Pagination should be removed
>>> from plain text RFCs only if there's rough consensus that
>>> nobody prints RFCs anymore - not because a few individuals
>>> think that nobody should do that, or think that they're in
>>> a position to dictate how RFCs are used. 
>>> 
>>> Keith
>>> 
>>> 
>> 
> 
>