Minutes_ MathML Core Meeting 29 Sept._ 2025

09/29/2025 MathML Core Meeting-core#296<https://github.com/w3c/mathml-core/issues/296>
Attendees

  *   Brian Kardell
  *   Louis Maher
  *   David Carlisle
  *   Simon Pieters
  *   Bruce Miller
  *   Eri Pérez
  *   Murray Sargent
  *   Moritz Schubotz
Regrets

  *   Neil Soiffer
Action Items
2. Scribing / IRC

  *   Should we just use Zakim?
·        TPAC?

  *
  *   WhatNot / WhatUp
ACTION: Louis will communicate with blind IRC users to learn how scribing can be done efficiently with IRC. This scribing path must include a spell checker and should have a grammar checker.
2.1. Should all MathML elements really be potential hyperlinks / match :visited?<https://github.com/w3c/mathml-core/issues/142>
CONSENSUS: DC writes: straw poll on call 29th Sept that there is a mild preference for mrow/href but that the group will accept an a if the browser vendors prefer to implement <a>.
7. Describe how to layout DOM elements violating HTML5 model #57<https://github.com/w3c/mathml-core/issues/57>
It was closed but we ran into the second bullet point this week and it is still an issue
CONSENSUS: EP writes: Follow Firefox behavior of rendering text nodes inside MathML elements (for example, <mrow>text</mrow> should render the text). I will send a PR to the spec.
8. Should math-style, math-shift and math-depth have a discrete animation type? #293<https://github.com/w3c/mathml-core/issues/293>
EP: The proposed solution is to update the spec with these animation types: Which are discrete for style and shift, and as an integer for math test, then merge the tests and do the implementations in the browsers.
CONSENSUS: DC wrote: Propose to update spec with animation types as discrete for math-style and math-shift and integer for math-depth. @eerii will make PR for spec and tests.
Agenda Items
1. Announcements/Updates
MuS: I have a package that translates Unicode Math into MathML, And Brian wrote a nifty post about how to that can be used with Markdown. I am trying to get that running.
From Brian Kardell to everyone: https://bkardell.com/blog/11tyMath.html
2. Scribing / IRC

  *   should we just use Zakim?
·        TPAC?

  *
  *   WhatNot / WhatUp
BK: IRC scribes can communicate efficiently with GitHub issues which help ensure that issues receive comments.
BK would prefer using the IRC-based scribing.
ACTION: Louis will communicate with blind IRC users to learn how scribing can be done efficiently with IRC. This scribing path must include a spell checker and should have a grammar checker.
2.1. Should all MathML elements really be potential hyperlinks / match :visited?<https://github.com/w3c/mathml-core/issues/142>
BK: SVG let you put href on anything.
BK: We are down to whether we should introduce an a element or mrow to get href.
BK: Sanitizers will not complain about an a.
SP: From Mozilla's point of view, we prefer introducing an a element to MathML so that it is consistent between web languages that browsers support.
MuS: Wouldn't it be easy to write a polyfill to translate the hrefs into, a's?
DC: Yes, but MathML lives in other places than just the web.
MoS: It would be good to decide something.
DC: So, you've either got to use the mrow with an href, or an <a> with an href.
BK: At the end of the day this group feels that we will accept either answer.
SP writes: Our preference from Gecko's POV is to use MathML a's for links.
BK is on the <a> side.
CONSENSUS: DC writes: straw poll on call 29th Sept that there is a mild preference for mrow/href but that the group will accept an a if the browser vendors prefer to implement <a>.
BK: Once we have a decision, we can put it into MathML core.
3. Trimming of whitespace in attribute #165<https://github.com/w3c/mathml-core/issues/165>
4. [Simplify whitespace collapsing/trimming for text nodes, Support
in Core? #272](https://github.com/w3c/mathml-core/issues/272)
5. mtable support #125<https://github.com/w3c/mathml-core/issues/125>
6. CSS interoperability with vertical-align #273<https://github.com/w3c/mathml-core/issues/273>
7. Describe how to layout DOM elements violating HTML5 model #57<https://github.com/w3c/mathml-core/issues/57>
It was closed but we ran into the second bullet point this week and it is still an issue
EP: The second point says that text nodes inside MathML nodes do not accept text. In WebKit and Chromium, they do not render tests, but in Firefox, they do, and we could not find anything on the MathML core spec that says one way or the other, and it seems to have been closed without a resolution for this specific point. The issue has been reopened. What do we want to do?
BK: Is there a reason not to render it?
EP: I am not sure, but I do not think there should be a reason not to.
BK: It is almost less important that what we do makes sense than that we all do the same thing.
DC: We should not put an extra burden on the implementers.
DC: We should do something that makes sense then document it.
BK Does it show in Chrome?
EP: No, it shows on Firefox, and it does not in WebKit and Chromium.
BK: If you put an mrow with text in it, it does not show in Chromium.
BK: It has to be a bug. The default is to show content not hide it.
CONSENSUS: EP writes: Follow Firefox behavior of rendering text nodes inside MathML elements (for example, <mrow>text</mrow> should render the text). I will send a PR to the spec.
8. Should math-style, math-shift and math-depth have a discrete animation type? #293<https://github.com/w3c/mathml-core/issues/293>
EP: The proposed solution is to update the spec with these animation types: Which are discrete for style and shift, and as an integer for math test, then merge the tests and do the implementations in the browsers.
CONSENSUS: DC wrote: Propose to update spec with animation types as discrete for math-style and math-shift and integer for math-depth. @eerii will make PR for spec and tests.
Zoom Meeting Summary
Summary
The meeting began with calendar coordination and welcoming new participants before addressing challenges with IRC-based scribing and discussing MathML link implementation options. The team evaluated different approaches for MathML linking using either A or MRO elements, ultimately leaning towards the A element while acknowledging browser vendor preferences. The group concluded by resolving animation types for match style and shift, and addressing text rendering inconsistencies across browsers, with specific tasks assigned to team members for implementation and documentation.
IRC Scribing Challenges and Alternatives
The group discussed challenges with using IRC-based scribing, with Louis expressing difficulties in using it effectively while Brian noted its advantages for linking GitHub issues. They considered revisiting IRC scribing at a later time, with Louis open to exploring more accessible IRC clients and potentially seeking advice from visually impaired IRC users. The conversation concluded with Brian mentioning a need to address an ongoing issue about links in their working group, which he planned to discuss with Simon and Ana Van Kestron.
MathML Link Implementation Options
The team discussed two options for implementing links in MathML: using an A element or an MRO element with an href attribute. While there were arguments for both approaches, the group leaned towards using the A element due to browser support and sanitizer compatibility. Simon noted that a WebKit bug preventing A element display in Safari was recently fixed. The team acknowledged that either choice would be acceptable, but they emphasized that getting links working was more important than the specific element used. Bruce raised concerns about adding complexity to the markup, while David and others discussed the potential impact on existing implementations and generators.
MathML Linking Element Decision
The group discussed the choice between using the MRO or A element for linking in MathML, with a slight preference expressed for MRO due to its easier implementation path. While browsers like Mozilla prefer the A element, the group agreed to accept the browser vendors' decision if they choose A, as this would finally provide an answer to the long-standing question. The discussion concluded with plans to document this in the relevant GitHub issue and potentially include it in MathML Core, pending final browser consensus.
Animation Types for Math-style, math-shift, and math-depth
The team discussed animation types for math-style, math-shift, and math-depth, resolving to use discrete animation for style and shift, and integer animation for math-depth. Eri will update the spec with these animation types, merge tests, and submit PRs for implementation in Firefox and Chromium. The team agreed on this resolution after Eri explained the technical details and proposed solutions.
MathML Text Rendering Consensus
The team discussed an issue with text rendering in MathML nodes, where different browsers (WebKit/Chromium vs Firefox) handle text inside non-Mtext elements differently. They decided to follow Firefox's behavior and render the text, with Eri tasked to open an issue, modify the spec, and file browser bugs to implement this. The team also agreed to be more forgiving in error handling for unknown MathML elements, following the HTML5 default of showing content rather than hiding it.


Regards
Louis Maher
Phone: 713-444-7838
Email: ljmaher03@outlook.com