Minutes: MathML Full meeting 10 Nov, 2022

 Attendees:

   - Neil Soiffer
   - Louis Maher
   - David Carlisle
   - Deyan Ginev
   - Cary Supalo
   - Bert Bos
   - David Farmer
   - Steve Noble
   - Sam Dooley
   - Murray Sargent
   - Bruce Miller

<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.0-rc2#cp-md-0-regrets>
Regrets
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.0-rc2#cp-md-0-agenda>
Agenda
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.0-rc2#cp-md-0-1-announcements-updates-progress-reports>1.
Announcements/Updates/Progress reports

NS: There has been no progress on the core spec CR.

DC: will chair the intent meeting for next week.
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.0-rc2#cp-md-0-2-deyan-suggests-a-href-https-github-com-w3c-mathml-issues-428-deprecate-none-element-a->2.
Deyan suggests: Deprecate element <https://github.com/w3c/mathml/issues/428>

DG: We have an element named "none". MathML treats it as an empty mrow. DG
proposes to deprecate the "none" element from the core spec.

NS: "none" is used in elementary math to indicate an empty column. Mrow can
do the same thing.

DC: We should check with Fred about deprecating "none". If it is not used,
then we can deprecate it. Make it a core issue.

*ACTION* DG: should make deprecating "none" a core issue, and check with
Fred to see if this deprecation would break anything.

SN: We would have to change the elementary math text in the spec. Full spec
has elementary math. Publishers will have to write their texts differently.
We will have to have new authoring guidelines.

SN: is not sure how this would affect some of Pearson's work. He said that
this might change how some of their assessments are written.

SD: Does not think that Pearson ever got on board with using the elementary
math elements that are in presentation.

SD: uses "none" in some of his code.

MUS: In looking through his code, he sees that "none" is in his reader and
writer.

NS: uses "none" in some of his software. The changes would be easy.

MUS: His code does not use MathML. "None" gets interpreted as an empty
argument.

NS: Must except both "none" and an empty MRow.

DG: HTML5 falls back to an empty mrow when it finds something that it does
not recognize.
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.0-rc2#cp-md-0-3-neil-39-s-simplified-core-intent-list-suggestion-a-href-https-docs-google-com-spreadsheets-d-1eswou1k5nxbdlpvqapdoa9h-s8lg_qjn8fjh64g9izq-edit-gid-140910536-see-quot-minimal-core-quot-and-quot-defaults-quot-a->3.
Neil's simplified core intent list suggestion see "Minimal Core" and
"Defaults"
<https://docs.google.com/spreadsheets/d/1EsWou1K5nxBdLPvQapdoA9h-s8lg_qjn8fJH64g9izQ/edit#gid=140910536>

NS: was tasked with coming up with an alternative minimum list for core. He
came up with the proposal.

NS: He wanted to keep things simple and to minimize the cases when intent
is used. Good defaults can reduce the need for intent.

NS: There are at least three cases when things must be specified in core.

   1. When the AT speech depends on the types of the arguments.

DF: discussed some of NS's examples like the absolute value, and open and
closed intervals.

   1. AT speech differs according to the needs of the listener. A blind
   person has different needs than does a person who has a visual processing
   disorder like dyslexia. The blind person wants to know when a square root
   ends whereas a dyslexic person might find this information confusing.


   1. The speech isn’t covered by using an intent concept@hint and is
   otherwise a notation that should be in core.

NS: AT will have certain defaults. For "mi" "mn" and "mo" the AT can just
speak the contents. Also, things like square roots and cube roots should
have special defaults. The AT default for "sup" would be to just read the
contents.

NS: continued to summarize his proposal.

NS: Unicode characters should be self-voicing. AT should have a list of
Unicode characters it can speak without intent.

From David Farmer to Everyone: So, in f : A \right arrow B I can put
intent=“from” on the colon and intent=“to” on the arrow.

NS: discussed special cases that the AT needs to know how to speak like
trig functions such as "sin" and operations such as log.

NS: discussed how you handle large operators such as summation and
integration.

NS: also discussed how you would speak large tables, and small tables like
a two-by-two matrix. A single column or row vector would be spoken
differently.

NS: it would be very useful to allow generators to say I am an intent aware
generator, and therefore, if I did not put intent on anything I want you to
follow the defaults.

NS: These special cases would come from a publicly available table.

NS: Parentheses and brackets may need to be in core to describe how to
speak expressions that use nesting.

NS: So, the AT would be required to figure out whether units are spoken as
singular or plural.

BM: We are conflating two things that causes confusion. 1. You halve a list
of things that need special treatment. This is called the core list. Then
there's a separate question of what elements need intents and what elements
do not need intents. This is determined by defaults.

NS: wants simple defaults that AT and generators can agree on and will not
need intent.

NS: does not want to use intents until a person encounters higher level
math.

MUS: You do not give AT enough credit that AT can get through calculus.

DG: NS's proposal will make things difficult for higher mathematics. Higher
math would not be spoken correctly by defaults. We must disable defaults
for higher math.

NS: AT defaults to powers currently when describing superscripts. This is
not appropriate for tensors and other higher math expressions.

BM: Anyone who is generating MathML needs to know what the assumptions are,
so that they can block the default behavior.

NS: We will have a meeting next week (November 17) and DC will chair it. NS
will send out the agenda.

DC: should talk about how to make intent simple.