Minutes: MathML Full meeting, 18 Dec, 2025

 Attendees:

   - Neil Soiffer
   - Louis Maher
   - David Carlisle
   - Bert Bos
   - Deyan Ginev
   - Moritz Schubotz
   - Murray Sargent
   - Patrick Ion

<https://cryptpad.fr/#cp-md-0-regrets>Regrets
<https://cryptpad.fr/#cp-md-0-action-items>Action Items
<https://cryptpad.fr/#cp-md-0-2-review-latest-spec-changes-and-see-if-we-are-ready-to-approve-moving-to-cr->2.
Review latest spec changes and see if we are ready to approve moving to CR?

*ACTION:* DC will add an example of <a href> to the full spec issue #555
<https://github.com/w3c/mathml/issues/555>).

*ACTION:* Someone should make an example of a polyfill using a tab-stop and
an OnClick event.

*ACTION:* The full spec is complete enough for NS to ask other groups to
perform their horizontal review.
<https://cryptpad.fr/#cp-md-0-agenda>Agenda
<https://cryptpad.fr/#cp-md-0-1-announcements-updates-progress-reports>1.
Announcements/Updates/Progress reports

Our next meeting will be on January 8. 2026. Hopefully, we will be able to
vote on going to CR for the full spec at that meeting.

DC reported bugs in the system that makes MathML in pdf accessible. One of
the problems is that it can render only one page at a time. If you navigate
with H to move through headings, it tells you there's no more headings
because there's no more headings on that page. If you page down, the
document magically appears again. The system cannot hold the entire file in
memory.

DG: it's kind of a miscellaneous updates-shaped question. We got a request
On how to better display tables in HTML. This was for an Archive user.

DG: I was wondering if we have a preferred way of modernizing displayed
equations?

NS: You should use display = block.

DG: So the reason they were in the table is that they have an equation
number, and they're centered with the equation number floating to the side.

DC: So from LaTeX, we stopped generating tables, and we just generate a
single math and put the equation numbers in an mtable, even if it's only a
one line.

NS: The suggestion I would have is to add role equals presentation, then it
is not considered to be a table for accessibility.

From Deyan Ginev to everyone: MARDI portal and Wikidata are Moritz
projects, to be clear. So the "we" there addresses their teams.

From Deyan Ginev to everyone: Requests often reflect the backgrounds of the
requesters :-)

MuS: I was curious about MathJAX and the anchor element, so I went to the
MathJAX
GitHub website and search the issues, and there is an issue, 3295
<https://github.com/mathjax/MathJax/issues/3295>, where the author is
complaining about how the anchor, while supported in all the browsers, does
not work correctly in MathJAX.

MUS: So, maybe something's going to happen, and MathJAX is going to fix
that.

NS: Well, did anybody check which browsers <A> works with MathML? So, you
know, if you just used an A instead of an mrow? Should all MathML elements
really be potential hyperlinks / match
<https://github.com/w3c/mathml-core/issues/142>.
<https://cryptpad.fr/#cp-md-1-2-review-latest-spec-changes-and-see-if-we-are-ready-to-approve-moving-to-cr->2.
Review latest spec changes and see if we are ready to approve moving to CR?

There has been progress in NS' decimalPoint issue and in DG's case
insensitivity issue.

DC approved their pull requests. They still need some editorial work.

DC: The core spec appendix b.3 does not make sense. It should be fixed.
This will be worked in the new year.

DC: We have not resolved what to tell users to do to get an accent.

NS has just added Add <a> to spec <https://github.com/w3c/mathml/issues/555>
As agreed in the Dec 4, 2025 meeting, <a> should be added to the full spec.
It should be considered to be identical to mrow with the comment that in
core, the href attribute works.

DC: In full, <A> can be used wherever mrow is allowed. <A> is an alias for
mrow. section 3.3.1 in the full spec.

DG: It would be helpful to have an example with <A HREF> used.

*ACTION:* DC will add an example of <a href> to the full spec issue #555
<https://github.com/w3c/mathml/issues/555>.

From Deyan Ginev to everyone: href="#" exactly

*ACTION:* Someone should make an example of a polyfill using a tab-stop and
an OnClick event.

DC The reason for deprecating aligngroup is that we can use tables.

Where do mrows span and column span show up inside math?

DC: In block matrices.

MuS: Can the polyfill handle the baseline alignment for align group and
align mark? He was having trouble with a fraction.

NS: MathML full can do it.

DG: There are examples of aligning in Archive.

*ACTION:* The full spec is complete enough for NS to ask other groups to
perform their horizontal review.
<https://cryptpad.fr/#cp-md-0-zoom-intent-meeting-summary-12-18-2025>Zoom
Intent Meeting Summary 12/18/2025 <https://cryptpad.fr/#cp-md-0-summary>
Summary

The meeting focused on reviewing progress and updates on the MathML
specification, including discussions on pull requests, polyfills, and
ongoing issues. David shared updates on Firefox's incremental progress and
issues with PDF accessibility, while Moritz provided insights on the
Wikimedia Foundation's strategy for rolling out MathML features. The group
discussed the deprecation of the maligngroup in favor of tables, with David
explaining the complexities involved. Neil emphasized the need for examples
in the spec and highlighted the importance of baseline alignment in tables.
The team also addressed concerns about row and column spans, with Patrick
and Deyan providing real-world examples of their usage. The conversation
ended with plans for the next steps, including preparing pull requests and
initiating a horizontal review process for the specification.
<https://cryptpad.fr/#cp-md-0-software-development-challenges>Software
Development Challenges

The meeting began with casual conversation about end-of-year workloads.
Neil inquired about the status of documentation tasks, to which David
explained he had been prioritizing paid work over volunteer tasks.
<https://cryptpad.fr/#cp-md-0-firefox-accessibility-and-math-layout>Firefox
Accessibility and Math Layout

David reported progress on Firefox accessibility improvements, noting that
structural elements are now mostly fixed but issues remain with incremental
rendering and NVDA support. The team discussed refactoring display math
equations in HTML, with Neil suggesting using the 'display: block' style on
math elements and adding a 'role=presentation' attribute to avoid
accessibility issues. Deyan inquired about modernizing displayed equations,
and David shared that LaTeX now uses a single math element with equation
numbers in an mtable, suggesting CSS adjustments for layout control. Neil
and Deyan agreed on using divs as a non-patch solution for layout
constructs and noted the 'role=presentation' suggestion as a good approach.
<https://cryptpad.fr/#cp-md-0-project-restart-and-rendering-updates>Project
Restart and Rendering Updates

Moritz and Deyan discussed the status of a project that has been inactive
for many months but is not officially abandoned. Moritz provided an update
on the progress of MathML rendering, explaining that they are working on a
plan with the Wikimedia Foundation to phase out SVG rendering. The
community has concerns about the rendering quality, and Moritz proposed
using MathJax as an intermediate step. The Wikimedia Foundation is
considering testing MathJax to ensure it meets their standards. Moritz
expressed optimism that the project could be completed by mid-next year,
with the MathML feature being rolled out alongside a larger change in the
platform.
<https://cryptpad.fr/#cp-md-0-mathjax-anchor-tag-issue>MathJax Anchor Tag
Issue

The group discussed an issue with MathJax and anchor elements, where Ross
Moore reported that despite browser support, the anchor tag doesn't work
correctly in MathJax. David mentioned that Ross is using an HTML A inside
mtext, which is a different case than using A inside MathML. The team
agreed to track this issue, with David planning to add it to the meeting
minutes and potentially open a new issue for further investigation.
<https://cryptpad.fr/#cp-md-0-pull-request-review-and-updates>Pull Request
Review and Updates

The team reviewed progress on pull requests, discussing the approval and
merging of two pull requests related to decimal points and case
insensitivity. They identified an issue in Appendix B.3 of the core
specification that needs to be fixed in the new year. David introduced a
new pull request (558) that adds support for the A element, noting that it
acts the same as mrow but allows href attributes. Deyan suggested adding an
example of using A href, and David agreed to add one before merging the
pull request.
URI and maligngroup Updates

The team discussed two main topics: URI handling in browsers and the
deprecation of the malign group feature. For URI handling, they clarified
that while browsers restrict certain elements from using links due to
security concerns, a polyfill can be used to add click functionality.
Regarding the malign group deprecation, David explained that the feature
could be replaced by tables, though this requires complex implementation
details that need to be addressed in the specification. The team agreed to
simplify the documentation to indicate that table-based replacement is
possible, while acknowledging the complications involved, and David planned
to update the pull request accordingly.
<https://cryptpad.fr/#cp-md-0-mathml-table-features-discussion>MathML Table
Features Discussion

The group discussed MathML table features, particularly focusing on column
spans and baseline alignment. David agreed to make a pull request over the
weekend to add 'A' to the call, and Neil planned to contact other groups
about starting the horizontal review process in the new year. The team
acknowledged that while editorial fixes like character entities and
line-breaking examples were needed, these didn't affect the core review for
security, internationalization, and accessibility. They agreed to hold
their first meeting of the new year on January 8th to discuss moving the
specification into CR status.