RE: Math WG comments on latest CDF documents

Steve, and the CDF WG,

Thank you for your response to the Math WG's comments. Unfortunately, we
don't feel that our concerns are adequately addressed by it.

We'd be happy to contribute to a MathML-based profile based on the
compound-by-inclusion framework. However, MathML may also appear as a child
document by reference (via an <object>), and we don't feel that the current
draft provides the necessary support. While restricting consideration of
layout issues to scalable elements is fine for the SVG-centered profile
documents (WICD Full and WICD Mobile), it seems inappropriate for a general
compound document framework. Some brief discussion of layout negotiation for
objects which are not scalable should appear in the WICD Core document, or
perhaps even in the Document Object Model section of the CDR Framework
document.

Our specific suggestion is to include a specification like the following:
  "The Document Object Model for a child document SHOULD make available to
the parent methods to return the width, height and depth (or 'baseline
offset') of the child content."
Such a stipulation would codify handling of <object> that has been supported
already by some user agents, and has allowed scripting to provide reasonable
display of inline MathML. On the other hand, we see publication of these
recommendations without such a provision as implying a step backward.

We don't believe that leaving such considerations for a MathML-based profile
is the best course, as we don't believe they apply only to MathML. Any child
document which gives rise to text-like content needs the same sort of
support.

If you believe that a provision of this sort is beyond the scope of these
recommendations, then it seems that that scope excludes essential
interoperability requirements of MathML objects (and other text-like
objects). We feel that you should in this case remove mention of support for
MathML and examples of MathML from them for now, as in our opinion these are
currently misleading. In particular, the section delineating the scope of
the CDR Framework document includes the text:
  "While it is clearly meant to serve as the basis for integrating W3C's
family of XML formats within its Interaction Domain (e.g., CSS, MathML, ..."
We believe that it's misleading to imply that the Framework as currently
written is usable for a wide variety of languages, and specifically for
MathML. 

In any event, we ask that layout (size) negotiation for text-like child
documents be added as a formal requirement for the Compound Document by
Inclusion work. We would suggest that the CDR Framework document explicitly
state that automatic size negotiation between parent and child is not
currently supported by CDR but will be addressed in CDI; this negotiation
should then include access to the baseline of child content.

Thanks very much for your consideration.

Ron Ausbrooks on behalf of the Math Working Group

Received on Tuesday, 14 November 2006 06:43:36 UTC