Minutes: MathML Full meeting, 12 Oct, 2023

 Attendees:

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

<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-regrets>
Regrets

   - Cary Supalo

<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-agenda>
Agenda
<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-1-announcements-updates-progress-reports>1.
Announcements/Updates/Progress reports
<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-dg-interop-request>DG
interop request

DG: It went well. He Invited the social media world to give us a thumbs up.
We have about 60 votes.

interop issue is at:
https://github.com/web-platform-tests/interop/issues/556

*ACTION* NS asked us to vote for our issue.
<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-dc-spec-updates>DC
spec updates

DC: Discussed the sections he is editing while showing the results by
screen sharing.

NS: said that this section needs more writing.

MUS: This is slightly off topic, but you said that you were doing this with
MathCAT and Mathcat is written in Rust, and it wasn't obvious how to
integrate it with a web page.

NS: There is a web assembly version of Rust, And that's the one that
interfaces with the web page.
2.Core concept list updates
<https://w3c.github.io/mathml-docs/intent-core-concepts/>
<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-a-can-say-quot-sine-x-quot-or-quot-sine-of-x-quot-this-means-that-the-trig-functions-can-be-both-prefix-and-function-do-we-really-want-to-duplicate-the-24-entries-as-is-done-for-quot-transpose->a)
can say "sine x" or "sine of x". This means that the trig functions can be
both prefix and function. Do we really want to duplicate the 24 entries as
is done for "transpose?

NS: discussed the transpose example.

NS: There are two entries for transpose, one for a prefix approach
"transpose t", and one for a function approach "transpose of t".

NS was not sure what to do with the 24 trig entries in the table. Do we
have to list a prefix and function case for each trig entry?

DC: we could have a property saying it is a prefix or function.

NS: was thinking about having an array where he could list the prefix, or
function, or postfix cases.

DC: You should not have to list the two cases for each entry.

NS: You could generate the extra verbiage from the first line of the entry.

PL: The function property means virtually the same as the prefix property.
Prefix could imply function, and vice versa, and you should not have to
list the function entry.

You can make one a specialization of the other.

NS: Function could be spoken as prefix.

PL: The spec should say that the function is a form of prefix.

PL There is no reason to use prefix. We shall just have the function
property.

NS: We will switch these items to just functions.
<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-b-other-additions-to-the-list->b)
Other additions to the list?
<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-i-a-href-https-docs-google-com-spreadsheets-d-1eswou1k5nxbdlpvqapdoa9h-s8lg_qjn8fjh64g9izq-edit-gid-1358098730-deyan-39-s-original-spreadsheet-a->i.
Deyan's original spreadsheet
<https://docs.google.com/spreadsheets/d/1EsWou1K5nxBdLPvQapdoA9h-s8lg_qjn8fJH64g9izQ/edit#gid=1358098730>

NS: wanted to skim through some lists and see if any of the items should
not be in core.

NS thought that the piecewise entry should be removed.

PL: This table came from content MathML and piecewise was in content MathML.

DG: Isn't co-domain written without a dash? This dash is a bit strange.

NS: If you do not put in the dash, the screen reader will speak it as one
word.

DC: said that a quick Google search indicated that most people do not use a
dash.

*ACTION* NS will remove the dash in codomain.

We had a discussion on minus, unary-minus, plus and unary-plus. Do we need
two names for both minus and plus?

NS Do we need plus as a concept name? You would put the Unicode for plus
and we do not want to have a concept name for each Unicode character.

PL: You do not need a concept name for things that are pronounced correctly
as a single letter.

NS: Things are in core if AT has special ways of pronouncing it.

DC: You cannot do math without having plus, minus, times, and division in
your core list.

DC: I wouldn't make an exception to add every Unicode character but having
the core basic arithmetic operators in the core does not seem too radical.

NS: If you put plus in there but leave out other Unicode symbols you would
confuse people about your criteria for including things in core.

NS: You put something in core if AT has a distinct way of pronouncing it.

PL: Minus must be in core, or the minus would be spoken as dash. You must
have an intent to say minus.

PL: We must tell the AT that the ascii dash will be a minus.

NS Core does not deal with dashes properly.

NS: can see an argument for keeping times in core because you have
invisible times cases. He argues for dropping plus and keeping minus and
times.

PL: If you keep minus, you must keep plus to be symmetric.

PL: You must have divide because divide can be shown by a colon.

NS: Wants to have a part of the table where the names of characters
identifies the characters.

*ACTION* DC and NS will figure out ways to manage operators that are spoken
with the name used in Unicode.

PL We could flag questionable cases now and perhaps remove them later.

PL We could have one version of core with questionable concepts and one
without them.

NS: We want to simplify our spec, and having two cases would not do that.

NS: Several characters can give the same pronunciation, and a single
character can be spoken in diverse ways.

NS: Under sets, union, and intersection, many signs are Probably just
characters.

DG: The cartesian product is very much not a character.

DG: There's a famous post about 7 notations for Cartesian products with
parenthesis so that is a complicated concept.

DG: One of the notations is a character, but there are another 6.

PL: We Are not trying to cover all notations.

DG: So, in the flow of the current conversation, when Neil says something
is a character, it needs to be only a character if we're going to dismiss
it so quickly; otherwise, we should keep it as a concept.

NS does not have to do anything special for the Cartesian product.

NS What is the name for the exponential function. you write exp of
something. Its name is simply exponential function. We could pronounce it
as e of something or exp of something. It does not need intent.

DC: Intent may not be on the obvious character.

NS: What is the point of the Gamma function? There are many functions. why
worry about the Gamma function. It can be removed from the list.

NS: So, vector, matrix, determinant, transpose, vector product, scalar
product, and outer product, are just characters.

PL: Vector product for me was the same sign as Cartesian product. So, it's
useful to have one of these have intent.

PL: For example, scalar products are quite often used as a times sign.

DG: Sometimes you have the little arrows on top of variable names like Big
V with an arrow on top could be the vector V. But you also may have a list
of Points so A one comma 2 comma 3 could be a vector, and that is getting
towards what matrix is usually, which is property. That says, here's
tabulated data, that's a matrix, And that's different than a variable name
that happens to be a vector matrix.

NS: And you know that vectors are sometimes just bold X.

NS: I do not think we have vector as a property, and it should be.
<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-ii-a-href-https-docs-google-com-spreadsheets-d-1clpaiy9kx5k-67rg6rjsaxerdsb-_iymgzatkqjshvg-edit-gid-0-david-f-39-s-wish-list-a->ii.
David F's wish list
<https://docs.google.com/spreadsheets/d/1cLPaIy9kX5K-67RG6rjSAXErDSB-_iYmgZaTKQjShVg/edit#gid=0>
<https://sandbox.cryptpad.info/code/inner.html?ver=5.5.0-b#cp-md-0-iii-a-href-https-gist-github-com-dginev-825078ae316c32c312436f42061b3d05-deyan-39-s-physics-list-a->iii.
Deyan's Physics list
<https://gist.github.com/dginev/825078ae316c32c312436f42061b3d05>

and Chemistry list
<https://gist.github.com/dginev/ff7e6e090b79a0389fc2eff2b9961331>
iv: Neil's MathCAT intent list and MathPlayer's rules list (subject area
inferences <https://github.com/w3c/mathml/issues/476>, units
<https://github.com/w3c/mathml/issues/475>)

Didn't get to these -- will be part of next week's agenda
3. Revisiting unresolved differentiation issue (473)
<https://github.com/w3c/mathml/issues/473>

NS: We have four different notations for differentiation. do we lump them
all together as a derivative? We could use the name derivative with an
extra argument giving the degree of the derivative.

PL: Do we want to have several intents with several concept names?

DC: Can we get away with just putting the intense on the D, the prime, and
the dot, and letting somebody else sort out the rest?

NS: I had not thought about putting the intent on the actual characters.

BM: There are at least three potential arguments to this operation. The
thing you're taking the derivative of, the thing with respect to which the
derivative is being taken, and the number of times to apply the derivative.
All these things can be omitted and implicit.

DG: Who has the heaviest parsing burden, the generator, or the AT that must
figure it out. DG says to make the generator do the most work.

NS: David, can you detail your proposal as a comment?

We will bring up this list next week.