Re: Minutes: MathML Full meeting 5June, 2025

Resending with the correct date of the meeting...

> Attendees:
>
>    - Neil Soiffer
>    - Louis Maher
>    - David Carlisle
>    - Moritz Schubotz
>    - Murray Sargent
>    - Bruce Miller
>    - Deyan Ginev
>    - Paul Libbrecht
>
> <https://cryptpad.fr/#cp-md-0-action-items>Action Items
> <https://cryptpad.fr/#cp-md-0-1-announcements-updates-progress-reports>1.
> Announcements/Updates/Progress reports2. #247: Spec should specify what
> char to use for accents/lines <https://github.com/w3c/mathml/issues/247>
>
> *ACTION:* NS: We should say something about issue 247
> <https://github.com/w3c/mathml/issues/247> and move this issue along.
> 3. Round table: what do people think needs to be tackled in the next 4 - 8
> weeks, not counting spec writing and polyfill writing?
>
> *ACTION:* NS: Look to see how ARIA, or some other group, are using web
> platform tests, which may not be directly testing the browser.
>
> *ACTION:* If you think of a new issue put in this list. Begin charter
> writing in August.
> <https://cryptpad.fr/#cp-md-0-agenda>Agenda
> <https://cryptpad.fr/#cp-md-1-1-announcements-updates-progress-reports>1.
> Announcements/Updates/Progress reports
>
> MathML Core progressing to CR
>
> NS: I haven't seen any updates on Core's progress towards being a CR. I
> think it's a roughly two-week project.
>
> MoS: We had an announcement by the Wikimedia Foundation that they are
> going to switch to MathML as a default (
> https://phabricator.wikimedia.org/T271001).
>
> MoS: There was pushback from the community saying that the spacing was not
> perfect and other issues. The MathML rollout is being postponed.
>
> MoS wants the rollout to be fully compliant with MathML core.
>
> NS was thinking that MoS would want the group to push out more bug reports
> to the browsers to improve rendering.
>
> There was a discussion about spacing problems in browsers.
>
> MuS said Microsoft Word handled spaces more accurately.
>
> PL: We're progressing very slowly into the verification of our book, and
> I've been using standard speak-aloud functions, and I was surprised to see
> that many math symbols are not pronounced by browsers.
>
> NS: I think a lot of people don't realize that speech engines don't
> support a lot of Unicode characters.
>
> NS: MathCAT, and other programs, have a big dictionary of math symbols and
> how to speak them..
>
> DG: We have been using JavaScripts for short term solutions. The point I
> wanted to make is that the whole point of the standard is to avoid needing
> JS to render the math in the browser. So that JS is a stopgap until level 2
> is realized.
>
> <https://cryptpad.fr/#cp-md-1-2-a-href-https-github-com-w3c-mathml-issues-247-247-spec-should-specify-what-char-to-use-for-accents-lines-a->2.
> #247: Spec should specify what char to use for accents/lines
> <https://github.com/w3c/mathml/issues/247>
>
> (updates?)
>
> NS: Fred said that this is a font issue not a browser issue.
>
> NS: If one browser does one thing and another browser does another thing,
> then the spec should settle the issue.
>
> NS: There have been no recent comments on this issue.
>
> BM: We should say something about this. I do not care what we do, other
> than we should not be recommending combining characters Unless they're
> preceded by some sort of combining neutralizer, like a space. We should
> list the appropriate characters to use.
>
> *ACTION* NS: We should say something about issue 247
> <https://github.com/w3c/mathml/issues/247> and move this issue along.
>
> <https://cryptpad.fr/#cp-md-1-3-round-table-what-do-people-think-needs-to-be-tackled-in-the-next-4-8-weeks-not-counting-spec-writing-and-polyfill-writing->3.
> Round table: what do people think needs to be tackled in the next 4 - 8
> weeks, not counting spec writing and polyfill writing?
>
> NS: What should we focus on?
>
> NS: Previously we talked about how to get intent implemented.
>
> DG just wanted to get to CR. If he had any key issues, he would have made
> an issue for them.
>
> NS: DG and MoS were working on the open list.
>
> DG: Open list work should not hold up the spec.
>
> NS believes that implementations are extremely important.
>
> DG: When is our current charter ending?
>
> NS: Our current charter ends in October.
>
> DG: We have a need for a third charter.
>
> NS: We should focus on the third charter in August.
>
> PL is focused on web platform testing (WPT).
>
> PL is also interested in implementations.
>
> MuS: I'm quite interested in how we get these ambitious ideas actually
> working. We can stick it in through TeX Macros, which can be considered TeX
> control words.
>
> NS: You're talking about getting intense into MathML? Yeah, okay.
>
> MuS: I added a way to do it with Unicode Math, it's a little ugly. You
> know, Unicode Math tries to be a mathematical notation. This is clearly
> markup, not a mathematical notation. It sort of works.
>
> MuS: Another thing that I think we can probably just simply dismiss is
> that we're not going to get intents into braille. There are some semantics,
> for example, in Nemeth Braille, like it knows what to choose what a
> binomial coefficient is. So that's well-defined, but when it comes to other
> things, uh, I don't think there's a really clean way of getting intents
> into Braille. So that means that we wouldn't be able to round trip Intense
> Through Braille. Maybe we can just forget about that.
>
> MuS: How can we get intents into different systems? MuS suggested a
> WYSIWYG approach. He did that in some of his work.
>
> MuS is hopeful about AI because it can tell us what the document means.
>
> NS We must have at least two implementations of consuming intent.
> Currently we have MuS's and NS's work.
>
> NS: I don't know, David, whether your LaTeX project is going to count, and
> I hope that the Digital Library of Mathematical Functions (DLMF) is also
> one that counts from the front-end side of it.
>
> DC has a face-to-face LaTex meeting from this Friday to Monday with
> intents and MathML as major topics.
>
> NS: So the approach you're thinking of is a macro that allows someone to
> add intent, whereas another approach would be to add a style file?
>
> DC discussed their invisible times macro which inserts intents. This does
> not work in the case of using an associated file. The process of adding the
> intents makes it drop the content which is taken from the pdf stream.
>
> DC: After the LaTeX work we should work on the polyfills.
>
> DC: There's the technical things for how to do it, but, you know, for a
> lot of things, it's not clear what you should do, like in the case of the
> accents.
>
> NS: I hope you get your macro in there that allows intent, and then I
> would like to see on top of that, a style file that takes advantage of that
> and defines some more standard macros so that they're easier for people to
> generate.
>
> NS to LM: It would be good to have not just the spec, but a user practice
> guide.
>
> MoS: Yeah, I want to stress again that I think the most important thing is
> to really finish the core spec and the polyfills. Then we can finish the
> open dictionary with DG.
>
> MoS: There is my idea, of this MathML implementation guidelines, so that
> we have some guidelines to move from LaTeX to MathML.
>
> MoS: And then there is this giant project I started with Howard and Bruce
> many years ago to generate all the LaTeX macros for specific concepts, and
> it would be good if this effort could be continued, and in the end,
> generate intent.
>
> BM: There are some issues that need clarification like recommended
> characters for accents, and what's supposed to happen with stretchiness in
> various situations.
>
> BM: The full spec seems on-track though we have a large backlog of getting
> intents implemented.
>
> NS: We need to get the spec written and the polyfills written. He is
> concerned with the implementations in the browsers.
>
> NS: It's still not in a position where I could go to the MathJax folks and
> say, you need to have a MathML output, because the rendering's good, and
> you can save yourself a lot of time.
>
> NS: The rendering is not good. It should not be left totally to the
> browsers. There's so many things that just don't look quite right, some
> spacing issues, and what symbols do you use for the characters.
>
> NS: We should write web platform tests that demonstrate the rendering
> problems.
>
> NS: A lot of tests were added to check bug fixes, but not a lot of test to
> demonstrate implementation failures.
>
> NS: There is someone responsible for vetting tests to see if they are
> necessary.
>
> NS had considered writing many test cases of super and subscripts.
>
> NS does not see people turning down tests.
>
> NS: You could write a lot of tests if they were unique.
>
> PL: We may test other things than just the browsers.
>
> NS: The ARIA group has been adding some web platform tests. They aren't
> testing so much what the browser is doing, but sort of what The screen
> reader does in the browser, and I really probably should look to see how
> they end up doing this. Maybe it's just whether the browser is exposing The
> ARIA in the right way.
>
> *ACTION:* NS: Look to see how ARIA, or some other group, are using web
> platform tests, which may not be directly testing the browser.
>
> NS: Does anyone want to comment on the issues we have discussed?
>
> DC: We should push the polyfills more, changing the spec may be less
> important.
>
> DC: We should sweep through the spec to see what is not in core, and see
> if there is a polyfill for it, and, if not, write an issue for that.
>
> DC: I was going to raise an issue about the use of the Shadow DOM which is
> totally masking all my wonderful links. I'm not sure why we need the shadow
> DOM. It is useful. It would allow you to keep the visible DOM structure and
> put stuff behind it, But you don't end up with a mathematical structure,
> you end up with an empty MathML element.
>
> NS: For one of our meetings, we should focus on how polyfills should work.
> There are several ways to write polyfills.
>
> *ACTION:* If you think of a new issue put in this list. Begin charter
> writing in August.
>
>