Minutes: MathML Full meeting 19 Jan, 2023

 Attendees:

   - Neil Soiffer
   - Louis Maher
   - David Carlisle
   - Dennis Müller
   - Murray Sargent
   - Paul Libbrecht
   - Bruce Miller
   - Sam Dooley
   - David Farmer

<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-regrets>
Regrets

   - Bert Bos
   - Deyan Ginev

<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-agenda>
Agenda
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-1-announcements-updates-progress-reports>1.
Announcements/Updates/Progress reports

NS: The current version of Microsoft Edge now supports MathML. The Opera
browser will soon support MathML.

Missing from minutes? Did we resolve Intent for "mtable" constructs #337
<https://github.com/w3c/mathml/issues/337>

LM: Did not find any reference to deciding on issue 337 in the minutes.
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-2-charter-discussion-what-needs-to-be-changed->2.
Charter discussion: what needs to be changed?

NS: We need a new charter.

NS: reviewed the charter using his shared screen.

PL: We have not made any progress with internationalization for MathML. We
have not been able to coordinate with other groups on internationalization.
Internationalization could be a motivator for the new charter.

NS: Internationalization is in scope.

NS: reviewed the bullet points stating the charter goals.

DC: We can say that we have made satisfactory progress on core. The goal of
the next charter could be core level two.

DC: We may not have a working draft by May.

NS: considered developing a test suite of about twenty examples showing how
intent can be used to generate speech.

NS: has some personal tests, but they are not in the form that could be
referred to in the specification.

NS: Do we have an example that shows how AriaLabel interacts with intent?
No.

DC: We should not commit to MathML 5. The goal of the new charter should be
to complete work on MathML 4.

DC: We should leave the current charter scope unchanged. We should say that
we will finish the work described in the current charter.

PL: Should we add a study for a multi-language survey? This might be a
scope item for our next charter.

DC: We want time to finish our MathML 4 work. MathML 5 would take another
three years at least.

*ACTION* NS will have a discussion with BK to get the core group to
consider the next charter.

NS: We have many items that say need spec. We have plenty of work without
changing the scope.

NS: We should drop search because no one is working on it.

NS: Intent may help content markup which might help searchability, but
intent is for AT speech generation.

PL: We could see if public search engines would work on intent.

PL: A sight specific search might work on intent. NIST might have such a
search engine.

BM: NIST search engines do do some indexing.

NS: Should we leave search in the specification goals?

DM: talked about a search engine.

NS: Can we say that we might get a search capability?

BM's and DG's group are working on search.

NS: In an expression, if there is no intent, NS will make guesses as to the
meaning of the expression. He considers this better than just reading what
is in the text.

BM: This search capability would not happen in our charter's lifetime.

DC's group is focusing on pdf tagging. In their next release, math
expressions can be tagged with the text in the TeX source.

NS: If the TeX had a macro in it, DC's method would grab the TeX text and
put it into the pdf document as a tag.

DC: There is a problem in detecting where an expression ends.

NS: This does not affect our scope.

NS: It is a matter of getting the authors to use the appropriate macros.

NS: Can your macros create intent?

DF: Each math topic area does not require many macros because you have
subject area assumptions to guide the intent translation.

DM: People working in a limited mathematical area will have a limited
number of macros which they use. Because math has many subject areas, you
will have to work with many macros.

DF: If you work in an area where you know all the macros, then you will use
them. If you are working in an area which has hundreds of macros, you may
not use them because of the difficulty of remembering, or finding, the
correct macro.

DM: A search engine might solve this problem.

NS: Do we have anything new to add to the scope?

SW: Do we have mechanisms, to align equations with one another, when the
equations are embedded in text?

NS: We do have an IndentAlign attribute, which uses an ID, to help align
equations with one another.

NS: He does not know how well the IndentAlign attribute works.

NS: In core level two, line breaking will be discussed. hopefully, it will
agree with MathML 3. MathJax might have implemented this alignment.

SW: We should add alignment as part of the scope.

SW: What is the best way to add equation numbers?

NS: mlabeltr was taken out of core.

MUS: There is a way to add an equation number by using an array. This is
the single expression mode. It is not used in the paragraph mode.

MUS: In his technique, Linebreaking is not implemented. It works for simple
cases. The math layout engine has two levels: single equation mode and
paragraph mode. This technique was cut in 2007.

NS: CSS has counters. Aligning equation numbers with their corresponding
equations is difficult. Can CSS capture this alignment?

DC: Use an HTML table.

SW: Is this something that should be addressed? To make a table to add
equation numbers tends to defeat accessibility. You also want to reference
figures and tables with numbers. People might like to click on an equation
number and jump to the equation.

NS: Is there a pure CSS method to make equation numbers without using a
table?

DC: It is difficult to have equation numbers in multi-line expressions.

DC: If you have a multi-line equation, you want to find the baseline and
float the number to the edge of the page. CSS has problems with this.

NS: Can you use an ID to position the equation number?

DC: You can do simple cases.

*ACTION* Add equation numbers to the MathML core discussion, unless CSS
already does this.

NS: Malign is used for vertical alignment, and not horizontal alignment.

BM: This is not math at all. It is something at the document level and not
at the math level. This needs rethinking to see how math zones interact
with the document.

NS: next considered out-of-scope items.

NS: Leave content alone except for changes that might affect presentation
MathML.

SD: Do not make changes that would break previous versions of content
MathML.

DC: The scope is to finish MathML 4. Do not change anything.

NS: We should have a deliverable to have a first draft of core level 2.

NS: We can do a cr next month. We just need to work a little on intent.

*ACTION* DC: said that we should readopt the entity spec.

NS: People should read over the charter and be ready to talk about it next
week.

NS: will start writing the new charter.