[rfc-i] v3imp *Overview* Needed improvements to xml2rfc v3 canonical format

Sean Leonard dev+ietf at seantek.com
Fri Jan 23 01:00:39 PST 2015


Colleagues,

At IETF 90 in Toronto, I spoke up during the RFC session to express the 
need for several enhancements to the new XML-based canonical format. I 
was asked to put together a complete list of these needs, and to present 
them to the list.

Since that time I have been gathering a corpus of work that demonstrates 
needs, not just "wants". This e-mail series is intended to lay out these 
needs. Since several months have passed since July, some of these needs 
have been addressed in the new xml2rfc v3 drafts. For that I am 
grateful. I tried to comb through the mailing list(s) and latest drafts, 
so if I missed some things that have been addressed, please just 
politely point that out.

This e-mail is the *Overview*. Please comment on the specific 
enhancements in the subsequent e-mails, each of which will focus on 
particular items.

***
Major Improvements:

#1 Control over spacing and line-breaking
  respect authorial choices regarding spacing (preserve spacing), 
including line breaks, in spec-text
  includes "vspace", "br", "nobr", "wbr"

#2 Control over paginated output
  "keep with next paragraph if possible"
  "keep lines together"
  "optional page breaks"
  etc.

#3 Verbatim text
  variables, named elements, sample output, sample input, productions, 
code snippets, etc. in spec-text
  possibly monospace--specific presentation depends on output

#4 Ruby text
  <ruby> or Unicode interlinear annotation

#5 Tag figs with filenames, Internet message data
   media types + parameters,
   (computer) language

#6 Byte preservation for figs
   Attach "files" to RFC; exact octets recoverable

#7 Absolute figure alignment
   control spacing so that figures are not always indented with the 
spec-text: suggest abs-left, abs-center, abs-right (absolute alignment 
to page/medium)

***
Minor Improvements:

#8 Fragment tagging on sourcecode
   distinguish between whole (compilable) source code units, and 
fragments of source code

#9 Is "cbor" a type?
   probably mean CBOR diagnostic notation; CBOR is a binary format
   (compare "asn.1" vs. "ber/cer/der")

***
Tool Requests:

#A Convert to PDF with a quality tool
  e.g., include CSS pagination stuff, bookmarks, hyperlinks
  IETF license to Prince XML

#B Fix @page margin left header footer rules in xml2rfc

***

Thank you for reading.

Sean



More information about the rfc-interest mailing list