Iljitsch van Beijnum wrote:
> My apologies for the subject line. I'm very disappointed that the silent
> majority of draft authors isn't speaking up. I can't imagine that the
> vast majority of draft authors has absolutely no problems with XML2RFC.
I use the RFC 2629 format for all my drafts and cannot imagine using
something else. The xml2rfc tool is not perfect but does a good job
to convert in html or text.
I do not think that the RFC 2629 format should be the only mandatory
format, but I think that the upload tool should at least give a
warning when an RFC 2629 document will not permit to recreate the
text version uploaded at the same time.
I wrote an (yet unpublished) tool to convert RFC 2629 documents to
mobi document, so they can be used on an Amazon Kindle. The plan
was to take the list of all I-Ds published as RFC 2629 documents
before the deadline, convert them and let people install them on
their Kindle, e.g. to read on the trip to the IETF meeting.
Unfortunately most of the I-D uploaded in RFC 2629 form cannot be
used to generated a useful document (mobi or text), for two main
reasons: 1) the date element is not fully filled, so the date in the
final output changes and, 2) the references are not resolved (i.e.
there is <?rfc include='file'?> elements in the document), so the
final output changes when the documents referenced change.
The problem is that there is at the same time some high level
components (include, unspecified dates) mixed with low level
components in the RFC 2629 format. The low level components are
part of the subset that makes possible to rebuild the various output
formats at anytime and that, IMO, should be enforced at upload time.
The high level components are what makes the editing process
easier. I use a superset of RFC 2629 for my documents (e.g. I use
xinclude instead of the include PI) and a script to convert it to
the low-level subset described above. This script uses the docName
in the high level document to generate the filename of the low level
document, which is very useful when using source control management.
The script also resolves all the xinclude references, removes the
comments and adds the current date. The process is then something
like this:
turn-uri.xml
--> draft-ietf-behave-turn-uri-02.xml --> upload
--> draft-ietf-behave-turn-uri-02.txt --> upload
--> draft-ietf-behave-turn-uri-02.mobi
--
Marc Petit-Huguenin
Home: marc@xxxxxxxxxxxxxxxxxx
Work: petithug@xxxxxxx
_______________________________________________
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf