Minutes: MathML Full meeting 22 jan, 2026

 Attendees:

   - Neil Soiffer
   - Louis Maher
   - David Carlisle
   - Paul Libbrecht
   - Bruce Miller
   - Deyan Ginev
   - Murray Sargent
   - Moritz Schubotz
   - Bert Bos
   - 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-the-maligngroup-malignmark-pr>2.
Review the maligngroup/malignmark PR

*ACTION:* MuS should review section 3.5, in the full spec, because word is
generating malign syntax.

*ACTION:* People should read section 3.5 and see if it makes sense.

*ACTION:* We should discuss the accent mark situation at the January 26
core meeting.

*ACTION:* NS will work on the 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

There is a core meeting on Monday, January 26, 2026.

DC: JAWS reading MathML in pdf
https://latex3.github.io/tagging-project/documentation/wtpdf/fulldoc#firefox-af-jaws

DC The pdf reader error which caused the document to be unreadable if the
document had a table of contents was fixed.

The functioning pdf readers are Acrobat, Firefox, and Foxit.
<https://cryptpad.fr/#cp-md-1-2-review-the-maligngroup-malignmark-pr>2.
Review the maligngroup/malignmark PR

PR now merged can be reviewed in place at
https://www.w3.org/TR/mathml4/#presm_malign

DC was having problems getting this PR to TR space. It was failing its
validation with content MathML.

DC got it to tr space.

*ACTION:* MuS should review section 3.5, in the full spec, because word is
generating malign syntax.

*ACTION:* People should read section 3.5 and see if it makes sense.

The next thing we considered was core appendix B.3
<https://w3c.github.io/mathml-core/#combining-character-equivalences> This
table gives mappings between spacing and non-spacing characters when used
in MathML accent constructs.

The conclusion is that there is no character in core that will generate an
accent mark for all fonts and for all browsers.

NS: The people interested in mathematics should decide on how the accent
mark should be generated.

NS believes that it will be easier to get people to change their fonts
rather than their browsers.

DC: This table showed so many problems that people got depressed.

PI How can I write Newton's law without a double dot?

DG: We do have some browser implementers in the group; However, we do not
have a font developer in the group.

From Deyan to everyone: if we fork something and call it "the MathML
Polyfill" font would be a nice name

*ACTION:* We should discuss the accent mark situation at the January 26
core meeting.

NS: We will not have an intent meeting next week unless there is a specific
need for the meeting.

*ACTION:* NS will work on the horizontal review.
<https://cryptpad.fr/#cp-md-0-zoom-intent-meeting-summary-1-22-2026>Zoom
Intent Meeting Summary 1/22/2026 <https://cryptpad.fr/#cp-md-0-summary>
Summary

The meeting focused on discussing challenges with MathML font support and
character rendering across different browsers and fonts. David presented
updates on Firefox's new tagged PDF with MathML support, which will be
released on February 24th, and shared a comprehensive table showing varying
levels of support for combining vs. non-combining characters across
different fonts. The group debated whether to recommend using non-spacing
or combining characters, with Bruce suggesting they need to make a clear
specification choice. Murray proposed using non-spacing marks but suggested
a polyfill that would translate from non-spacing to combining characters.
The team agreed to discuss this further at the upcoming Core meeting on
Monday, where they will also determine their next meeting schedule based on
whether any urgent issues arise during the week.
<https://cryptpad.fr/#cp-md-0-team-updates-and-technical-challenges>Team
Updates and Technical Challenges

The meeting began with a discussion about pending changes, where David
confirmed he had completed most of the suggested modifications except for
one that he had pushed back on. There will be a core meeting on Monday.
<https://cryptpad.fr/#cp-md-0-firefox-mathml-pdf-update>Firefox MathML PDF
Update

David announced that Firefox now supports tagged PDF with MathML, which
will be available in stable release on February 24th, and noted that JAWS
can now read these PDFs for the first time. Moritz and Deyan reported
progress on an open list project, though they indicated it would be several
weeks before they could provide a demo due to performance issues and bot
traffic. The conversation ended with a brief mention of David's PR for the
Am aligned group.
<https://cryptpad.fr/#cp-md-0-mathml-pr-validation-updates>MathML PR
Validation Updates

David and Neil discussed updates to a PR, focusing on issues with Content
MathML validation and examples. David explained that he disabled two
examples due to validation failures and updated images in the PR. They also
discussed changes to text regarding table alignment and spanning
rows/columns, with Neil expressing doubt about the claims of the JavaScript
polyfill's capabilities. David noted that the text about Maligngroups was
left unchanged due to compatibility concerns, though he acknowledged its
awkwardness. Neil suggested clarifying that the first maligngroup should be
present, and David agreed to make this change. They also briefly touched on
the historical incorrect usage of "vertically aligned" versus "horizontally
aligned" in MathML.
<https://cryptpad.fr/#cp-md-0-mathematical-alignment-and-character-tables>Mathematical
Alignment and Character Tables

David discussed changes to a mathematical alignment in section 3.5 and
mentioned a pull request for review on Monday. He also addressed issues
with Appendix B.3, which contained inconsistent and redundant tables of
combining and non-combining characters. David suggested deleting the
redundant tables and replacing them with a single, reorganized table, but
expressed doubt that this change would be accepted by Core due to CSS
limitations. He proposed the need for a polyfill to handle character
selection.
<https://cryptpad.fr/#cp-md-0-mathml-character-encoding-challenges>MathML
Character Encoding Challenges

The team discussed challenges with character encoding and font support in
MathML, particularly regarding combining characters and ASCII
representations. David and Neil reviewed a test table that highlighted
inconsistencies across different browsers, with Firefox being one of the
few that properly normalizes characters. Bruce suggested creating a list of
feasible solutions that avoid using combining characters, though the team
acknowledged that current implementations have significant limitations,
with only a few characters working consistently across browsers.
<https://cryptpad.fr/#cp-md-0-font-compatibility-and-stretchy-characters>Font
Compatibility and Stretchy Characters

David, Bruce, and Neil discussed font compatibility issues, particularly
with stretchy characters and combining characters. They noted that some
fonts, like XITS and IBM Plex, do not support stretchy characters, while
others, like Latin Modern and Computer Modern, can be updated. The group
considered specifying which combining characters to use and creating a
polyfill for browsers that don't support them. They also discussed the
possibility of working with font developers to improve support for these
characters, though they acknowledged that some popular fonts, like those
from tech giants, are unlikely to change.
<https://cryptpad.fr/#cp-md-0-character-rendering-challenges-in-browsers>Character
Rendering Challenges in Browsers

The group discussed challenges with character rendering in different
browsers and fonts, particularly focusing on non-combining and combining
characters. They explored the possibility of implementing a polyfill to
automatically convert non-combining characters to their combining
counterparts, though concerns were raised about which characters to
prioritize. Murray suggested supporting non-spacing marks in XML while
recommending the use of combining marks in fonts, while David emphasized
that using combining characters for accents doesn't align with the Unicode
model. The discussion concluded with Deyan noting the absence of font
implementers in the core group, suggesting that having a reference font
could simplify solutions.
<https://cryptpad.fr/#cp-md-0-math-fonts-discussion-challenges>Math Fonts
Discussion Challenges

The group discussed challenges with math fonts, particularly regarding
stretchy brackets and accent characters. David noted that while some fonts
work well, there's no universally good solution, and the team agreed to
discuss this further at the upcoming CORE meeting next Monday. David
mentioned he would be away for LaTeX activities from Friday through Monday.