As we (WRI and Dave) discussed on Jan 24th, I will maintain position papers at WRI which describe the HTML-Math proposal as we (the members of w3c-math-erb@w3.org) see it at any given time, including our design goals and our approaches and specific proposals for meeting them, and also the aspects which require further discussion.
The initial versions of these papers are now available at http://....
The wording of the documents is intended to be that of a finished, precise specification of a standard, so it's in the present tense even about unfinished things, and it may be a bit pedantic in its level of detail.
In order to make clear which areas have reached a consensus, and which issues require further discussion, I'll try to put everything which is not something we've already discussed and essentially agreed upon inside square brackets ( [ ]s ). The material inside [ ]s will sometimes include (or refer to) my own proposals which have not yet been discussed, and sometimes my "editorial comments" about which issues need further analysis and/or discussion, perhaps including direct questions for our consideration. I need feedback on anything inside []s, and my goal is to move every such thing out of them (by discussing the issues identified, or by agreeing on the proposals given, or on something else) as fast as is reasonable. In general, I may assume that anything not commented on for a week has been agreed to.
The documents are organized as one contents page with links to each of several documents, each covering one issue or area. I'm hoping that as we fill in and agree upon the details, no more than one more level of links will be needed, to "appendices" giving complete descriptions of details proposals (like character sets, complete lists of default parsing rules, etc).
(There is a certain amount of redundancy between the areas, because I felt it was important to cover all the aspects of each area in one place. But in general, any one issue will be definitively discussed in only document, with the others referring to it.)
The documents themselves are text files. I'll probably convert them to HTML sometime.
I'll try to update these documents to reflect the results of our discussions with a time lag of only a few days. I'll summarize every update to the mailing list, and each document will tell when it was last updated. After each update, we'll discuss whether I'm right in identifying the areas of consensus and the issues remaining to be discussed, and we'll discuss those remaining issues.
Bruce (bruce@wolfram.com)