Re: More liberal draft formatting standards required

[John C Klensin <john-ietf at jck.com>]

While I use xml2rfc and do so reasonably happily, a few comments 
below...

--On Sunday, June 28, 2009 6:33 PM +0200 Iljitsch van Beijnum 
<iljitsch at muada.com> wrote:

> Hi,
>
> XML2RFC isn't working for me.
>
> For instance, We are now required to use boilerplate that the
> "official" version of XML2RFC doesn't recognize so it's
> ...
> I used to write drafts by hand sometimes in the past, but this
> is also very hard, because today's tools just don't have any
> notion of hard line endings, let alone with spaces at the
> beginning of all lines and hard page breaks (at places that
> make no sense in an A4 world, too).

Gee.  I've had few of those problems when I edit with emacs or 
its clones.  I guess that doesn't qualify as "today's tools" in 
your book but, if so, your issue is much broader than xml2rfc.

> This is getting worse because the checks done on IDs upon
> submission are getting stricter and stricter.
> ...
> As such, I want to see the following:
>
> - the latest boilerplate is published in an easy to copy&paste
> format

Agree that it is a problem if it is not readily available in 
that form.  But the Trustees seem to be too busy creating new 
policies.

> - drafts may omit page breaks

While I've stopped noticing whether the non-requirement is being 
enforced, to my knowledge, page breaks have _never_ been 
required in I-Ds.  The problem is that the Secretariat likes (or 
has been told to) published a page count as part of the I-D 
announcement and that they can't figure out the page count 
without page breaks.  On would think that they could get a good 
approximation by scanning the document counting lines as well as 
looking for page breaks to count, noticing that there are no 
page breaks, and dividing the line count by 56, but, for some 
reason, that has been considered too hard for a decade or so.

> - drafts may omit indentation and hard line breaks

Indentation should not (ever) have been a problem with I-Ds.

Line breaks are because a situation in which some documents have 
them and others interfere with a different set of reading (and 
printing) tools.  Just as you want to be able to say "I want to 
be able to submit drafts without using xml2rfc to format them", 
I want to be able to say "I don't want my drafts to require 
processing through some display formatter before I can read them 
or have a discussion with someone else that includes references 
to "Line M of Paragraph 2 of Section 5.2".   I also note that 
some of us are very dependent of diffs and the like which also 
depend on well-defined lines.  So your soft line break 
requirement is the one that strikes me as completely 
unreasonable.

> - no requirements for reference formats

To the best of my knowledge, this has never been a requirement 
for I-Ds and the posting tools don't check it.  I'd be happier 
if xml2rfc permitted me to write something like:

<reference anchor="FooBar" status="later" />

and have it generate
  [FooBar] ... to be supplied ...

and/or
<reference anchor="FooBaz" status="incomplete">
  <t>whatever notes I want to write to myself, potentially with
   embedded cref elements </t>
</reference>

I'd be happy to have either of the above generate warnings, but 
they should not prevent the document from compiling.  On the 
other hand, I've figured out how to get around that sort of 
thing, so I've never considered it very important.

> Note that this is for drafts in general. If the RFC editor
> wishes to impose stricter formatting rules I can live with
> that.

Good.  Because they certainly will.  Note also that there is a 
potential difference between a working draft that you are 
sharing with friends or a WG and incrementally developing and a 
draft that the IESG is willing to look at or put out for Last 
Call.  Certainly it would be reasonable to subject the latter to 
requirements that more closely resemble the RFC ones than it is 
for the former.

> ...

   john