Minutes: MathML Full Meeting 14 August, 2025

 Attendees:

   - Neil Soiffer
   - Louis Maher
   - David Carlisle
   - Deyan Ginev
   - Murray Sargent
   - Bruce Miller
   - Paul Libbrecht
   - Bert Bos

<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-a-href-https-www-w3-org-math-documents-charter2023-html-charter-extension-a->2.
Charter Extension <https://www.w3.org/Math/Documents/Charter2023.html>

Pick an amount of time and verify that we will not change the charter
during that time.

*Consensus:* The group agreed to request a charter extension to the end of
2026.

*ACTION:* NS will construct a rough timeline as part of our extension
request which he will send to BB and the group. BB will send it to the W3C.
<https://cryptpad.fr/#cp-md-0-3-progress-on-spec-writing->3. Progress on
spec-writing?
None :-(
<https://cryptpad.fr/#cp-md-0-4-a-href-https-github-com-w3c-mathml-issues-524-524-naming-the-arguments-to-intent-new-property-needed-a->4.
#524: Naming the arguments to intent: new property needed?
<https://github.com/w3c/mathml/issues/524>

*Consensus:* NS: The consensus is that authors will have the option to use
the argument names during navigation by use of the ":argname" property. So
we're adding the use-argname property To our known properties. The other
part of the consensus is that the concept table
<https://w3c.github.io/mathml-docs/intent-core-concepts/>, both core and
open, should have a column that allows someone for core, that suggests
names for the slots, and for open, allows somebody to propose names for the
slots along with their proposed concept name.
<https://cryptpad.fr/#cp-md-0-agenda>Agenda
<https://cryptpad.fr/#cp-md-0-1-announcements-updates-progress-reports>1.
Announcements/Updates/Progress reports

NS experienced problems when he tried to export LibreOffice files
containing MathML to EPUB.

DG followed up this issue and received a detailed answer from Regina
Henschel, a LibreOffice contributor. There are available workflows for
MathML interoperability and continuing support.

DG: Nothing has happened for the mid-August date with my intent open list
work. My new goal is mid-December.

PL: There is a very nice surprise which happens on the Mac. If you have
MathML on your text on office for the Mac, and you paste it into Excel, you
see beautiful math, but if you hit enter to keep the Math, then the math
gets linearized.

Mus: I had a version that I never checked in where you could paste math
into Excel, and the cells would display beautifully because they used
regedit, but it never shipped.
<https://cryptpad.fr/#cp-md-1-2-a-href-https-www-w3-org-math-documents-charter2023-html-charter-extension-a->2.
Charter Extension <https://www.w3.org/Math/Documents/Charter2023.html>

Pick an amount of time and verify that we will not change the charter
during that time.

BB: We can pick whatever time we want even up to two years. The W3C would
prefer you ask for a longer time rather than coming back for a second
extension. We cannot change the charter during the extension.

NS reviewed the scope, deliverables, and timeline of the charter.

DC: We have to get core level one finished.

NS: We already have some things listed for core level two.

NS: The draft says there are open questions about links and line breaking.

DG: Can we make a draft without fully specifying line breaking?

NS: Yes.

NS: We can start refining some of our next tags.

NS to BB: We should come up internally with what we expect the delivery
dates to be.

NS: Success criteria. In order to advance to PR, each normative
specification expected to have at least two independent interoperable
implementations of every defined feature, where interoperable can be
verified by passing open test suites and/or with two or more
implementations interoperating with each other.

NS: We may have to figure out what an open test suite is.

NS: I think one of the things we just need to do is write a document that
has everything that we are claiming, like here's a fraction, now here's a
fraction with an intent on it, and walk through all the options for intents.

NS: Each specification should contain sections detailing security and
privacy implications.

NS: Do we need to modify the charter? group says no.

NS does not see us producing a candidate recommendation until late November.

Igalia has this work item where they're working on making all the browsers
compatible, and that might take a year.

NS: twelve months for us might be too short because Igalia may run into
issues our group must handle.

NS: The key date to me is when we go to CR for MathML4, and I see that
happening in December.

DC says we should get a fourteen-month extension so that we end at the end
of 2026.

*Consensus:* The group agreed to request a charter extension to the end of
2026.

*ACTION:* NS will construct a rough timeline as part of our extension
request which he will send to the group and BB. BB will send it to the W3C.
<https://cryptpad.fr/#cp-md-1-3-progress-on-spec-writing->3. Progress on
spec-writing?
<https://cryptpad.fr/#cp-md-1-4-a-href-https-github-com-w3c-mathml-issues-524-524-naming-the-arguments-to-intent-new-property-needed-a->4.
#524: Naming the arguments to intent: new property needed?
<https://github.com/w3c/mathml/issues/524>

NS: We are naming the arguments of intents.

NS: I like David's notion, let's just use the names of the arguments.

NS: For some functions, arguments do not have names. For example, the
argument of the absolute value does not have a name.

DC: Most functions have generic arguments like A, B, and C. These arguments
do not have natural names.

Pl: Some function arguments have names that people do not use such as
mantissa and root.

NS: Some of the specialized argument names can be untranslatable into other
languages.

NS: There is no French word for an exponent, so his son's French teacher
says, and there is not English word for the argument of the absolute value.

NS: If there is no word in core for an argument, how can you translate it?

NS: Our core concept list
<https://w3c.github.io/mathml-docs/intent-core-concepts/> needs an
additional column suggesting what an argument can be called. The core list
name would not be a required term, but just a suggestion. The open terms
may also need a naming suggestion column.

NS: I do not know what to call the terms of the binomial coefficient.

MuS: You can call the terms top element and bottom element, or upper
element and lower element.

From Deyan Ginev to everyone: "outcomes" and "possibilities" are used by
MathWorld: https://mathworld.wolfram.com/BinomialCoefficient.html

BM: Most of the time you do not want unusual argument names to be used. Any
scheme used to assign names in the intent could be confusing. For the
special cases where it is desirable to have a special name, put the name in
the dictionary.

PL: In some languages, the top and bottom terms of the binomial coefficient
are reversed. From time to time, in special cases, specify the terms in the
dictionary.

NS: In summary, if the author really wants to give a name to the arguments
and have them used as a navigational aid, then they actually put :argname,
and that would override… The AT's decision, else the AT would fall back
into the dictionary.

*Consensus:* NS: The consensus is that authors will have the option to use
the argument names during navigation by use of the ":argname" property. So
we're adding the use-argname property To our known properties. The other
part of the consensus is that the concept table
<https://w3c.github.io/mathml-docs/intent-core-concepts/>, both core and
open, should have a column that allows someone for core, that suggests
names for the slots, and for open, allows somebody to propose names for the
slots along with their proposed concept name.

PL: We should leave it to the author not the specification.

DC: This issue gives the author control. If we put it in the dictionary,
then the author does not have control.
<https://cryptpad.fr/#cp-md-0-zoom-math-wg-intent-meeting-summary-2025-08-14->Zoom
Math WG Intent Meeting Summary (2025-08-14)
<https://cryptpad.fr/#cp-md-0-quick-recap>Quick recap

The team discussed technical challenges around AI-generated summaries and
translation issues, while also addressing MathML support problems in
various software platforms. They reviewed their charter extension process
and discussed progress on MathML specifications, including the development
of test suites and the creation of a public working draft for MathML Core
Level 2. The group focused on improving mathematical argument naming
conventions and documentation practices, while also planning to schedule
writing sessions to address long-standing tickets and complete
specifications before the end of the year.
<https://cryptpad.fr/#cp-md-0-next-steps>Next steps

   - Working Group: Aim for Candidate Recommendation for MathML 4 by late
   November/December 2025
   - Working Group: Complete remaining writing tasks for MathML 4
   specification before Christmas break
   - Bert: Prepare the charter extension request without editing the
   charter itself
   - Neil: Update the timeline for the charter extension request and send
   it to Bert
   - Working Group: Internally develop a list of updated expected dates for
   deliverables
   - Deyan: Continue working on the intent open contribution lists with a
   new target date of mid-December
   - Working Group: Create an open test suite for MathML 4 specification
   features
   - Working Group: Prepare documentation for testing all MathML features
   with intent annotations
   - Working Group: Consider organizing a dedicated 2-hour writing session
   call to complete pending documentation
   - David: Write content for open tickets he's responsible for
   - Working Group: Prepare a first public working draft of MathML Core
   Level 2 after finalizing Core Level 1
   - Working Group: Prepare a first working draft of MathML 5 that includes
   open issues and next tags
   - Bert: Forward the updated timeline to Philippe for the charter
   extension
   - Neil: Write something to Bert that he'll forward to Philippe for the
   charter extension
   - Working Group: Add a column to the Core specification table indicating
   suggested names for arguments
   - David: Update the core concepts table to include a column for
   suggested argument names
   - David: Add a "needs spec update" for the argument naming property in
   the intent attribute
   - Deyan: Look into Libre Office's MathML support issues
   - Deyan and Moritz: Include a column for argument names in the open
   concepts table they are working on
   - Neil: Incorporate Murray's suggested terms for binomial coefficients
   into MathCAT

<https://cryptpad.fr/#cp-md-0-summary>Summary
<https://cryptpad.fr/#cp-md-0-ai-translation-and-mathml-challenges>AI,
Translation, and MathML Challenges

The group discussed AI's ability to generate summaries from transcripts,
with Deyan explaining that the process involves navigating a
high-dimensional probability space similar to human thought. Neil shared
his experience with Norwegian and Swedish translation challenges,
particularly regarding neuter words and number pronunciations. The team
also addressed issues with LibreOffice's MathML support, noting that while
PDF exports display math properly, EPUB exports do not include math at all,
leading to a decision for Deyan to investigate this further. Paul reported
a surprising discovery on Mac Office where MathML text can be pasted into
Excel, but the math is then linearized upon pressing return.
<https://cryptpad.fr/#cp-md-0-charter-extension-and-mathml-progress>Charter
Extension and MathML Progress

The team discussed extending their charter, with Bert confirming they can
request a duration of up to 2 years, though Neil prefers a shorter period.
They reviewed the current charter scope, which includes developing
recommendations, test suites, and clarifications for web platform
technologies, with David noting they've made progress on most items. The
team agreed to create a first public working draft for MathML Core Level 2,
which will include open issues and proposed changes, while keeping the
existing charter unchanged during the extension period.
<https://cryptpad.fr/#cp-md-0-mathml-development-and-accessibility>MathML
Development and Accessibility

The meeting focused on the development and accessibility of MathML,
including the creation of a repository for intent values and properties
beyond those defined in the MathML 4 recommendation. David mentioned
starting the intent cycle for content MathML to presentation MathML
conversion, though this was paused due to a pivot towards accessibility.
Neil emphasized the need for at least two independent interoperable
implementations of each feature to advance to PR, and discussed the
importance of open test suites. The group also touched on the need for
testing plans for each specification, security, and privacy implications,
and coordination with external organizations like the WHAT Working Group.
<https://cryptpad.fr/#cp-md-0-charter-extension-and-work-progress>Charter
Extension and Work Progress

The group discussed extending their charter, agreeing to request an
extension until the end of 2026. Neil will update the timeline and send it
to Bert, who will forward it to Philippe for approval. They also discussed
the progress of their current work, including MathML 4 candidate
recommendation expected in late November or December, and the need to
complete writing tasks before then. The group acknowledged that some
tickets have been open for a long time and agreed to schedule a 2-hour
writing session to address this.
<https://cryptpad.fr/#cp-md-0-mathematical-argument-naming-conventions>Mathematical
Argument Naming Conventions

The team discussed naming conventions for mathematical arguments, with
David proposing to use argument names directly while Neil suggested adding
a nudge for macro writers to consider these names. They explored challenges
with naming conventions, particularly for functions without natural names,
and discussed the need for argument naming in core concepts. Murray
contributed examples from matrices and binomial coefficients, suggesting
alternative terms like "top" element" and "bottom element" to improve
clarity and readability. Neil acknowledged the importance of these
contributions and planned to incorporate them into MathCAT.
<https://cryptpad.fr/#cp-md-0-argument-name-handling-in-navigation>Argument
Name Handling in Navigation

The group discussed how to handle argument names in navigation and
documentation. They agreed that authors should have the option to specify
argument names using the "use argname" property, which would override the
default ATs behavior. They also decided to add a column to the core and
open concept tables that suggests names for slots, though these suggestions
would not be mandatory. David will update the specification to reflect
these changes. The conversation ended with a brief discussion about their
charter and goals for the coming year.