On Tue April 5 2005 15:30, Alex Rousskov wrote:
On Wed, 2005/03/02 (MST), <blilly(_at_)erols(_dot_)com> wrote:
I've suggested (via Reply-To) discussion on the IETF list.
Thanks a lot for reviewing and commenting on the draft!
I am preparing a revision to address Last Call comments.
I am not on the IETF list anymore (too much noise), so I am CCing Tools
discussion list instead. Please feel free to forward elsewhere.
I'm copying the IETF list for closure.
It seems odd that there is no provision for upload of nroff source
(RFC 2223) mentioned in sections 7 and 8 of the draft.
The motivation for uploading XML sources is that they are used by tools
and humans processing submitted drafts.
Likewise for nroff source.
For example, RFC Editor is often
using authors' XML sources.
While I have no data to either confirm or refute that assertion, RFC
2223 section 3 and the draft successor to that document both explicitly
state that the RFC Editor uses nroff.
Yes, but speaking from personal experience, that doesn't mean they'll use
_your_ nroff sources. I stopped bothering to send in my nroff sources when I
found out that the RFC Editor wasn't using it (this despite having tried to use
the same macro package in the same way, and having switched from a much better
system than nroff in the first place in order to be RFC Editor friendly).
Shortly after finding out about this I switched to using xml2rfc, in part
because Marshall spent considerable time aligning its nroff output with the RFC
Editor's needs. And some time after that it became possible to send the RFC
Editor the XML source and have them use it. I haven't looked back since.
We expect such uses to grow once XML sources
are easily available. In fact, the submission tool itself is expected to
extract useful metadata from XML sources.
Exactly. And powerful tools to assist in this sort of extraction are readily
I suspect that similar metadata could be extracted from nroff source,
at least if a suitable macro package (e.g. as described in
draft-lilly-using-troff) is used.
Bruce, with all due respect, the effort you have expended on developing this
seems to me to be headed in the wrong direction. IMO the place we are at with
xml2rfc and the RFC Editor's acceptance of the format is much better than
anything can ever get from any scheme that is based on nroff as the primary
What would be the motivation for uploading nroff sources?
In addition to extraction of metadata,
o nroff is used by the RFC-Editor (RFC 2223 section 3); keeping the
same source format from initial draft through RFC production can ease
the workload for authors and the RFC Editor
See above. Use of xml2rfc accomplishes this goal. Use of nroff does not.
o automatic generation of plain text, PostScript, PDF, HTML (including
line diagrams, tables, data formats, etc.), preserving page layout,
from single source
All possible with the xml2rc format.
o (if a suitable macro package is used) no need to upload boilerplate;
provided that the (IETF copy of the) macro package is maintained,
up-to-date boilerplate can be generated automatically
Xml2rfc has been realigned each time the boilerplate requirements have changed.
This has proved to be tremendously convenient.
o ability for authors w/o access to formatting tools to upload easily-
produced document source which can be used to produce a formatted
IMO neither nroff source nor XML source qualify as "easily produced". But the
number of tools for producing and vetting XML is large and growing. (I
personally use something called "Exchanger XML Editor".) I don't think the same
can be said for nroff.
Ietf mailing list