A variety of XMLspec RFEs

I have been trying to save off XMLspec requests for enhancement, and here's 
what I've got, in mildly organized form.  I'll try to do some analysis in 
followup messages (I realize it's a bit cryptic).

Markup/processing
=================

A priority
==========

*Add blockquote.

*Offer a <loc> element that reproduces the content as the generated
href attribute (if no href was supplied in the XML).

*Improve the setup of termdefs in a deflist scenario

*Scrap langusage and use xml:lang instead

*Don't set defaults on rowspan and colspan

*Support emph in code examples

*Distinguish normative vs. non-normative citations

*Graphic needs alt attribute to be required

*Table should have summary attribute

*MathML additions

- Talk to David Carlisle; point to his tools in appendix
- Expand eg to allow tech.class
- Expand sub and sup to allow at least var, and maybe tech.class
- Allow var inside code and eg
- Consider allowing tech.class in many more places in general

*Make content on nt optional so it can pick up the lhs of the prod it
points to.

*Make constraintnotes a class so they can be parameterized.  Do the
same for their content.

*Make issue text parameterized.

*Top-level attribute to indicate whether this is the normative version
of a spec or a non-normative one (e.g., a translation).

*Scope of availability of links (e.g., member-only vs. public vs. team-
only).

*Add more XLink attributes

*Add inline graphics

- Use case: needed for math

*In header, capture deadline for comments and where to send them

*Harmonize when to use href on bibl and/or titleref

*Consider further changes to the latestloc region

- Look at James Clark's handling
- See Jonathan Marsh's XSLT snippet for how to handle both cases

*Add value to occur attribute on arg to allow rep in addition to req

- See request from Oliver Becker

B priority
==========

*Support IETF requirements.

- See the RFC 2629 XML DTD: http://www.faqs.org/rfcs/rfc2629.html

- Documents can be RFC, FYI (For Your Information), BCP (Best Current
Practice), and STD (Standard).  (What about I-Ds?)

- An RFC may have a specific relationship to an earlier RFC: Updates or
Obsoletes.  In each case, the number of the earlier RFC must be
supplied.

- If it's an RFC, the subcategory can be one of: Standards Track, Best
Current Practice, Informational, or Experimental.  The status section
in these four cases is boilerplate.

- There's a canonical Introduction section, which is likely to need to
contain boilerplate.

- There's a canonical References section near the end.

- There's a canonical Security Considerations section near the end.

- There's a canonical Author's Address section at the end.

- There are authors instead of editors.

- Joint specifications might have special requirements, so
W3C/IETF/joint needs to be indicated.

- Need alternative graphic formats for ASCII output and other outputs

*Consider allowing role values to be RDF properties.

*Extraction of examples into "plain-text" chunks

- Need: hide me, highlight me, normal, ellipses, descriptive text
- See Ian notes and my notes

*DOM additions

- See http://www.w3.org/DOM/Group/drafts/pubtext/xmlspec-v21-dom.dtd

*XML Schema additions

*Add ability to "model" XML vocabularies

*Raw bibliography entries

*Build in XInclude capability for examples

*Use Dublin Core somehow for the header stuff

*Mark which productions are (what?????)


Documentation
=============

Make the documentation conform to the WCAG Add textual
descriptions of diagrams (e.g., the real content model etc.)

Upgrade documentation into Note form

Point to latest versions of tools, e.g. Arbortext doctype.

Add proper citation to SWeb documentation See CMSMQ's mail

Add best practices (e.g., breaking up chapters into entities) to
documentation

Point to XMetaL CSS customization for XMLspec

Point to Bob Stayton's LiveDTD tool

Harmonize the FPIs in the documentation and the actual references


Stylesheets and processing
==========================

Allow chunking and non-chunking output (can do this already?)

Get everything into CVS

Use background colors instead of tables for code examples

Extraction of examples into "plain-text" chunks (see above)

Make XMLspec formatting a comma tool

Produce a TOC for productions

Extract the productions for import into a tool (see above)

Extract issues into an issues list See mail in issue tracking
folder

Get Jonathan's IE-compatible stylesheet and put into CVS

Produce XSLT converter for any backwards incompatibilities
--
Eve Maler                                    +1 781 442 3190
Sun Microsystems XML Technology Center   eve.maler @ sun.com

Received on Wednesday, 17 October 2001 10:32:11 UTC