Minutes: MathML Full meeting 26, Jan 2023

 Attendees:

   - Neil Soiffer
   - Louis Maher
   - David Carlisle
   - Bruce Miller
   - Stephen Watt
   - Murray Sargent
   - Moritz Schubotz
   - Patrick Ion
   - Steve Noble
   - David Farmer
   - Paul Libbrech

<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-agenda>
Agenda
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-1-announcements-updates-progress-reports>1.
Announcements/Updates/Progress reports

LM: Said that using Zoom with the screen maximize greatly helps gain access
to Zoom controls. A user can maximize the Zoom meeting window for all
meetings by going to the Windows Zoom client, going to settings, general,
and tab to the item labeled "Enter full screen automatically when starting
or joining a meeting", and check that item. If you are in a Zoom meeting,
you can maximize the screen using the "alt f" command.

LM: also said: if you have Zoom menus open such as the attendance and chat
items, you can move rapidly between them by hitting f6.
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-2-charter-discussion-a-walk-through-with-some-quot-live-quot-changes->2.
Charter discussion: a walk through with some "live" changes?

NS: brought up the MathML WG charter and shared his screen. We then began
discussing different parts of the charter.

NS: Our current charter says we are overhauling MathML 3.

NS: We should have a candidate recommendation (CR) for MathML 4 by the time
our present charter is closed.

NS: Our next charter will produce a full recommendation which will include
implementations of MathML 4.

SD: We are progressing MathML 4 towards recommendation status.

NS: Should we drop search because we are not working on it?

MOS: I've been working on some search, but it's not something which will be
completed in the next year.

No one objected to removing search for the new charter.

SW: How long have chemical formulas been in the charter? There are many
types of chemical formulas with diagrammatic aspects.

NS: Chemistry was part of the original charter.

SW: My feeling is that we overstepped on that one.

NS: We have talked about chemistry in the "isa" discussions.

NS: Chemistry can be thought of as notational things that fit in MathML.

SW: Our group may never work on describing structural formulas.

CS: If we do not consider the highly diagram oriented formulas, are we
totally taking chemistry out of scope?

SN: We should not say that we are only working on MathML for mathematics.
MathML can be used for many other things. We should still leave chemistry
in scope.

SN: MathML can be used for linear chemical equations.

SN: We should say that we will work on MathML notation which can be used in
chemistry.

PI: We only promised increased accessibility for chemistry. We do not say
how far we are going to get.

Under scope:

PL: wants to strengthen internationalization by consulting with the
international community.

SW: The internationalization scope item lets us also talk about
localization: about how different parts of the world describe the same
mathematical expressions differently.

NS: We are working on refinements to the mathematical markup and processing
model that enhances Accessibility.

SN: We will use the term deprecation, instead of removing, for items we
wish to retire.

MOS: One of our goals was to foster the adoption of content MathML.

NS: We do not want to talk about changing content MathML in this charter.
This is a subject for the MathML 5 effort. If we put changing content
MathML in our current scope, we would never finish in our allotted time.

MOS: When we are happy with intent, we might consider how to improve
content MathML.

DC: wants us to go to full recommendation by the fourth quarter of 2023.

NS: There are many implementations which need to be written for the full
recommendation. We will not get this done in 2023.

NS: said that BM's group has LaTeX producing MathML presentation core.

DC: It's depressing that we are forecasting taking 2 years to get from Cr
to rec.

DC wants us to get to CR by March and get to recommendation when the
charter runs out in May.

NS: Should we put out the end date in one year?

DC: We do not want to ask for another extension.

NS: We should say we are beginning to explore MathML 5. For example, we are
Collecting requirements for MathML 5.
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-3-more-intent-discussions>3.
More intent discussions
<https://sandbox.cryptpad.info/code/inner.html?ver=5.2.2-0#cp-md-0-a-a-href-https-github-com-w3c-mathml-issues-409-409-internationalization-a->a)
409 internationalization <https://github.com/w3c/mathml/issues/409>

NS: The core of this comes down to is intent an English-only specification
which is translated into other languages or is it intent meant to be spoken
the way it is in other languages.

NS: If you meant intent to be spoken in French or Bulgarian or Chinese,
then intent should be in French, Bulgarian or Chinese, and that's what gets
spoken.

NS: If you have an image in French, then the alt tag should be in French.

NS: If you think of a TeX to MathML converter, and somebody has \abs, and
it's going to get translated into absolute value in English, it also needs
to get translated into whatever the equivalent of absolute value is in
French, Chinese or Bulgarian.

DC: If intent is in the speaker's spoken language, then intent is there to
give mathematical structure.

NS: Division might be described with different words in the same language.
We will have some advisory pronunciation templates. We can use more words
to make things clear.

NS: In units, you don't say over, you say "per", for example miles per
gallon Or meters per second.

PL: We want to speak in a fashion that enhances accessibility.

PL: We would say 'absolute value' in the local language.

NS: How does AT know when something needs translating?

PL: Does intent contain language information?

PL: Intent should tell the AT if translation is required.

NS: Maybe we can always say you have to put a language tag on, it if it's
not English, but how do you know whether you need to do a translation or
not unless it matches something that you know about?

MUS: I'm just thinking about the way that we did math speak. We came out
with tokens, and then we have a library that's been localized into, I guess
they're about 18 languages now, we're thinking about raising it to 50
languages which take the tokens and speak the tokens in whatever language
is required. I don't think that's necessarily practical for MathML.

NS: So that you could have a very small core if everything is translated,
or you could have a large core if AT needs to know to do a translation.

DF: As someone who's going to take what the author provides and output
MathML, Am I supposed to know what the language is, or am I just supposed
to indicate the meaning of the author's intention? I don't see how it's
reasonable for me to do anything with what the language the reader wants to
hear.

NS: So, it's kind of bizarre that someone's written this textbook in
French, but the intent is not French. The intent should be translated.

DF: Okay, here's my answer. There's a textbook and there's a French
translation of this textbook. When I can convert that, I'm going to have to
figure out what are the French intent values? What's the French for a cross
product? What's the French for Cartesian product?

DF: Why can't AT do the translation?

NS: AT would need to know all the possible names that are generated so it
can do the translations. This is probably a large list..

NS: The alternative is that the TeX macro needs a French version that
specifies how to say abs in French.

DF: My French author is not going to go and figure out how to markup
MathML.

SW: Two thoughts. One is that we have lots of examples of HTML pages where
there are inclusions in other languages. I don't know what mechanisms exist
for code switching between languages for screen readers, but shouldn't we
be looking at that number one? Number two is that in various fields of
mathematics, we do not always translate all the words. For example, in
Hebrew one may keep the English vocabulary for graph theory but translate
the vocabulary for elementary mathematics. And so, I believe that we should
think very hard about this. One does not always translate English into
other languages or vice versa. We should investigate what mechanisms are
used elsewhere for "code switching" languages.

PL: Textbook publishers will want control over how things are said.

NS to MUS: Can we get the list of things that are translated. I'm not
asking for Microsoft to disclose all the translations they've done, what
I'm asking for is a list of all the things that you translate as an initial
starting point for Core.

MUS: I can certainly give you that list.

DC: By using underscores, we can force the AT to speak translated text.

DF: We need to know how to markup intent for a certain collection of
instructions. Do we need to get this for every single language?

NS: urges people to contribute to the internationalization issue between
meetings so that it can be resolved in a meeting.

NS: we also need to work on "isa" which has contention.

NS: to DF: Can you come up with a list of things that need to be
translated? MUS also has a list. Let us look at the overlap. This is the
list that AT should translate.

NS: We have a core meeting on Monday January 30, 2023.

NS: It takes at least two months to get the charter approved. We should
submit the new charter by the end of February 2023.

NS: We will also work on the charter next Thursday, February 2.