Minutes: MathML Full Meeting 19 Sept, 2024

 Attendees:

   - Neil Soiffer
   - Louis Maher
   - David Carlisle
   - Deyan Ginev
   - Patrick Ion
   - Murray Sargent
   - Bruce Miller
   - Moritz Schubotz
   - Bert Bos

<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-0-regrets>
Regrets
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-0-action-items>Action
Items

NS: We will next meet on Thursday, September 26, during TPAC.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-0-2-a-href-https-github-com-w3c-mathml-issues-504-table-property-clarification-and-addition-of-grid-504-a->2.
Table property clarification and addition of :grid (504)
<https://github.com/w3c/mathml/issues/504>?

*ACTION:* NS: We added a :structure property which tells the system to
simply read what is there issue 492
<https://github.com/w3c/mathml/issues/492>. This property needs to be added
to the list of core properties in the spec. DC put need spec update. NS
will update the spec.

*ACTION* DC: We voted for the new property to be called array or table.
Array won. Array will cause the mtable to be read by rows and columns,
unless "by columns" is present which would reverse the reading order.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-0-3-review-units-and-currency->3.
Review Units and Currency:

   - Units and Currency Property text
   <https://w3c.github.io/mathml-docs/intent-core-properties/#Units-and-currency>
   - include standardized currencies in core intents #490
   <https://github.com/w3c/mathml/issues/490>
   - List of units #475 <https://github.com/w3c/mathml/issues/475>

*RESOLUTION:* Issue 475 is an extensive list of units and will be helpful
for implementers. However, since we have a "units" property, we will not
need this list. Issue 475 was closed.

*RESOLUTION:* Since we have a currency property, issue 490 was closed.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-0-4-other-issues->4.
Other issues:

   - Remove/Deprecate subscriptshift/superscriptshift attributes of
   msubsup/msup/msub #27 <https://github.com/w3c/mathml/issues/27>

*ACTION:* DG: I would like to ask Frederick a question about CSS Aligned
items, and how applicable that is to MathML because CSS has some online
properties, but I don't know how applicable they are to these situations.

*RESOLUTION:* Leave subscriptshift/superscriptshift in.

*ACTION** NS: There is a need for a polyfill for
subscriptshift/superscriptshift because core does not currently implement
it.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-0-agenda>
Agenda
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-0-1-announcements-updates-progress-reports>1.
Announcements/Updates/Progress reports

   -

   TPAC is 23-27 Sept. Shall we meet next week during TPAC? NS: We will
   next meet on Thursday, September 26, during TPAC.
   -

   Spec update
   -

   maligngroup/malignmark comments issue #181
   <https://github.com/w3c/mathml/issues/181>. And glyph comments: issue 25
   Remove/Deprecate mglyph element <https://github.com/w3c/mathml/issues/25>
   ?

DC: Do not use maligngroup and mglyph. They are not in core. See the notes
for them #507 <https://github.com/w3c/mathml/issues/507>

Replace mglyph with img or a polyfill.

DC also said that none should be replaced by an empty mrow.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-1-2-a-href-https-github-com-w3c-mathml-issues-504-table-property-clarification-and-addition-of-grid-504-a->2.
Table property clarification and addition of :grid (504)
<https://github.com/w3c/mathml/issues/504>?

MuS: Matrix just means to me something that doesn't have any brackets or
parentheses or curly braces or whatever. And if you want to have those
guys, then you put them in explicitly, and speech just follows suit.

NS: I was doing some work in MathCAT and there is nothing to read a simple
table like:

a b c d

NS: I thought, perhaps matrix is a poor choice of a name because that
implies more than that's what it's used for.

NS: The name we use, like matrix, should imply reading row by row and
column by column unless the "by column" property should override the
reading order.

NS: We need to have matrix. I think we need to have determinant, and I
think we need to have grid or table or something.

DC: This depends on people. People's view is on having an explicit version
of default, I mean.

DC: Would having grid be any different from not having it?

NS: Well, it seems to me like, if we have ways of speaking something, we
should have properties that allow you to say this is the way I want to
speak it.

DG: Yeah, I remember we made it the property because it was much easier to
annotate compared to making it a concept, because then the roles would need
to be arguments.

DG: So, if it's a table or a grid, and I think this is like classic
conversation, and the HTML now has both right. So, HTML has a table
element, and it has CSS. Grids, and they're really the same thing. So, it's
extremely confusing. But they're synonyms.

PI: I think that matrix always used to be just the grid of values, and that
it's only when you decide that you'll have a matrix algebra by composing
these things in some way or other, that that you have something further.

PI: Matrix used to be the grid of values. There is more than one matrix
multiplication. discuss if matrix is associated with parentheses. When you
have a determinant you say it is a determinate before you read the numbers.

BM: Seconding PI: Matrix is a math object involved in linear algebra
whereas array is a neutral data structure with as many dimensions as you
need.

PI: I don't think that there's any sense in which the word matrix is
associated with parentheses or verticals for a determinant or whatever.
And, furthermore, you don't re read the determinant values or cell values
differently. You say the whole thing is a determinant, and that assumes
that there's some algebra that's done on the collection.

NS: But you know, if there are the vertical bars outside, you don't say the
matrix of you say the determinant of, I think, David at 1 point, said the
determinant of the matrix. That's how he would say it.

MuS Array is a reasonable substitution for matrix.

From Deyan Ginev to everyone: which also works fine for the "det(A)" form.

PI: What do people do to read higher order arrays some simple, 3 by 3 by 3
array written as an array of sub matrices.

DC: We only have a 2d layout. We can have M. Tables with M. Tables.

DG: We have most of these things solved. We need a neutral property for the
grid.

NS: Array reads rows and columns. For a table it would tell you how big it
is.

DG: Was there a property to read things how they are visually written?

DC: We've got a structured property here.

NS: This means do not assume any math for anything, just read what is there.

BM: That's kind of what we're haggling about for array.

BM: This is a way to provide a default speech output. Structure is
overridden by intent.

NS: Does that mean that if you put a structure, if you put parentheses
around a table, put a structure property on it. I should stop saying the
matrix for it.

DG: We should specify what structure does.

DC structure not in spec.

*ACTION:* NS: We added a :structure property which tells the system to
simply read what is there issue 492
<https://github.com/w3c/mathml/issues/492>. This property needs to be added
to the list of core properties and the spec. DC: put need spec update. NS
will update the spec.

NS: You would want to put structure up high in the description.

NS Structure literally read what is there.

NS would prefer literal to structure.

DC literal is already used so stick with structure.

MuS: I typed in mathematical matrix into Wikipedia, and I got a good
article on it, and all of the matrices that are in there are bracketed or
parenthesized. They never even mentioned the possibility of not having some
sort of bracket involved. So in that sense, I think that calling an
unbracketed mtable an array kind of makes sense.

DG: A property controls speech not expression rendering.

*ACTION* DC: We voted for the new property to be called array or table.
Array won. Array will cause the mtable to be read by rows and columns,
unless "by columns" is present which would reverse the reading order.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-1-3-review-units-and-currency->3.
Review Units and Currency:

   - Units and Currency Property text
   <https://w3c.github.io/mathml-docs/intent-core-properties/#Units-and-currency>
   - include standardized currencies in core intents #490
   <https://github.com/w3c/mathml/issues/490>
   - List of units #475 <https://github.com/w3c/mathml/issues/475>

*RESOLUTION:* Issue 475 is an extensive list of units and will be helpful
for implementers. However, since we have a "units" property, we will not
need this list. Issue 475 was closed.

*RESOLUTION:* Since we have a currency property, issue 490 was closed.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.9.0-1#cp-md-1-4-other-issues->4.
Other issues:

   - Remove/Deprecate subscriptshift/superscriptshift attributes of
   msubsup/msup/msub #27 <https://github.com/w3c/mathml/issues/27>

NS: These attributes are not in core. They should be in core. They need to
stay in MathML full.

NS to DG, BNM, and MuS: do any of you use these?

BM: No, but since TeX uses it, I might want to try.

NS: How would you do this in TeX?

DC; Well, there's 2 ways. One is, I mean the font metrics, both classic
tech fonts and open that fonts have got font parameters that say how far to
drop the subscript.

DC: The other way, which is easier and therefore gets done more is as it's
mentioned in the comment there from 2020 s, is just to use a strut. you
know y strut under subscript x.

NS: So in MathML we have no way of affecting the fonts directly.

DC is unsure.

NS: So, Bruce, do you just ignore struts in the conversion?

BM: Yeah, I'd have to run through some tests to see what exactly happens.

NS: If people are generating them, then we should keep them in core.

NS: Do we say if Safari or Firefox are using them?

NS could not find a reference saying that Safari or Firefox use them.

PI has a Mac to see if the shift is available on Safari.

NS: MDN says subscriptshift/superscriptshift are Implemented in Safari but
not in other browsers.

DC: I've assumed it's not in core. So even if it's implemented, It's likely
to get removed from browsers. But it probably doesn't hurt to leave them
in, does it? And just say, polyfill them.

*ACTION:* DG: I would like to ask Frederick a question about CSS Aligned
items, and how applicable that is to MathML because CSS has some online
properties, but I don't know how applicable they are to these situations.

DC: Leave them in, and add links to where the polyfills are.

DG wrote: A problem does exist for supporting chemical notations as they
want sub/superscripts always aligned. I do not currently have enough
practical experience with CSS Box Alignment
<https://developer.mozilla.org/en-US/docs/Web/CSS/CSS_box_alignment>, but I
was wondering during the call today whether that could be a long-term
device to accomplish the script alignments for chemistry. If other
participants here have thoughts on that idea (especially when it comes to
MathML Core), it would be interesting to hear. To clarify David's comment
above, the call resolved to keep the attributes in MathML Full, as there is
no clear alternative mechanism available today.

*RESOLUTION:* Leave subscriptshift/superscriptshift in.

*ACTION* NS: There is a need for a polyfill for
subscriptshift/superscriptshift because core does not currently implement
it.