meeting minutes from 28/5/19

 Once again, thanks to David Carlisle for taking notes during the meeting.
Also thanks to the several members of the chemistry group that joined the
call and gave us great input. The meeting was *not* recorded, so this is
the only record. Apologies.

*Brief* status report on core working group's progress

Neil: MathML Core is moving forward. Lots of issues have been identified
(see github <https://github.com/mathml-refresh/mathml/issues>).  Mostly
removing items that are "difficult" for browser implementers from core,
keeping in main spec. Aligning mathml attribute syntax with css where
possible (in full). There will be polyfills mapping from full spec to core.

Positive feedback from chrome developers so hopeful that it will be picked
up by Chrome, although no promises yet. The new Edge will be based on
Chomium, so good chance it will go in there. Also, Frederick did the
Firefox implementation, so changes likely will go in there. What happens
with safari/webkit is unknown.

Most issues marked with needing more tests and/or spec changes.

Daniel: what is planning for merging implementation to main branch of
chromium.

Neil: I believe it is planned to be nearer the end of implementation (Fred
or Rob would need to provide more detailed answer)
Distinguishing Chemistry from Math (from a Chemistry CG)

Neil: there is a chemistry CG (some on the call) Chemical Formulas use
similar constructs to math (sub/super scripts arrows, etc.) There are a few
small layout differences between math and chemistry. Want screen reader to
know it is chemistry not so that something like H20 is not pronounced “cap
H sub 2 cap 0” but pronounced “H 2 0” or maybe even “water”.

Neil: one possibility role="chemistry"

Murray: generally, elements don't use math italic

Neil: also positions of sub/superscripts are slightly different in
chemistry than in math

Jason: I'm not a chemist but wanted to clarify distinction is between
mathematics and chemistry, or whether it is between expressions involving
or not chemical systems, also use of units of measure.

Cary: Something else Jason and I talked about is dimensional analysis.
E.g., units need to be carried through, so need to pass between
mathematical and chemical notations in same equation

Dan: Ions with + and - distinguishing that between mathematical use of
these symbols

Neil: MathPlayer implemented speaking chemistry, Ions are not a problem
once you recognize that (for example) the H is Hydrogen not a math variable.

Jason: could have an attribute on <mi>, but would people refer general
indicator on each mi, or at level of math

Neil: for interpretation may be better to be higher up.  E.g., to change
pronunciation of a right arrow to “yields”.

Tzviya: easy to special case elements but when marking up novel compounds
very hard to have automatic rules and need author input to distinguish

Tzviya: not just pronunciation, but also layout and hyphenation and linking
to other compounds.

Dan: even in school level books hard to give alt texts for unnamed
constructs.

Jason: You would need to read the constituents if the construct has no
name, so more a rendering implementation rather than markup issue.

Cary: K means at least 4 different things in Chemistry, hard to get correct
reading. It could be lower case, capital, script, or roman. My screen
reader doesn’t tell me that normally.

Tzviya: many publishers in have a list of PUA characters of chemistry

DavidC: That was a bad thing for the early MathML versions and could be a
blocker for any future spec.

Murray: the UTC has a procedure for submitting to Unicode.

DavidC: TeX has a couple of chem packages. They probably push the limits of
what can be done with chem formulas and math editors. I will contact some
chemistry package authors for TeX to see what issues they address.
From David, here are links to the packages:
    siunitx:
http://mirrors.ctan.org/macros/latex/contrib/siunitx/siunitx.pdf
    mhchem:  http://mirrors.ctan.org/macros/latex/contrib/mhchem/mhchem.pdf
    chemformula:
http://mirrors.ctan.org/macros/latex/contrib/chemformula/chemformula_en.pdf



Neil: I will add individual issues to github as needed, also others can add
issues to repo.
    [These are issues #92
<https://github.com/mathml-refresh/mathml/issues/92> and #93
<https://github.com/mathml-refresh/mathml/issues/93>]

Murray: Maybe the Unicode sans serif math alphabets could be used to avoid
them being in italics.
Adding some semantic info to presentation MathML (#64
<https://github.com/mathml-refresh/mathml/issues/64>, #68
<https://github.com/mathml-refresh/mathml/issues/68>, #79
<https://github.com/mathml-refresh/mathml/issues/79>, #80
<https://github.com/mathml-refresh/mathml/issues/80>, #84
<https://github.com/mathml-refresh/mathml/issues/84>, ???) Neil: At a
recent hackathon I was at, I spoke with James Nurthen who co-chairs the
ARIA WG. He referenced [de]pub as a potential model for adding role
information for math. I was concerned that we would be adding too many
roles but James pointed out that dpub prefixes its roles.

Tzviya: I'm one of the authors of dpub-aria

Neil: all additional roles are prefixed dpub- so we could potentially add
roles with names math-...

Neil: roles could be used to distinguish different fraction-like notations

Tzviya: it is a lot of work to produce the roles, recommend starting with
the Aria Authoring Practices <https://www.w3.org/TR/wai-aria-practices-1.1/>
document
It is important that any role has a super class, dpub maps roles to an
existing role
this means that authoring content is complicated as rules on where each
role is valid.
Especially for hand authoring will get errors

Neil: as we are not placing roles on already defined elements in the ARIA
spec, we do not have some of the constraints that placing aria on html.

Tzviya: Someone would have to work on mapping this to accessibility API

Jason: Aria WG would probably insist on mapping to accessibility API but
they may be amenable to adding extensions for better support of Math.

Neil: we have a longstanding invitation to discuss this with Aria group,
but I have hesitated until we have proposal for what we need.

Tzviya: good article on mapping at accessibility
https://alistapart.com/article/semantics-to-screen-readers/

Neil: math is tricky/different because the words you use vary depending
upon the disability of the user. What is good for one person might be bad
for another.

Tzviya: encourage conversations with Aria WG sooner rather than later.

Charles: there is a personalization task force (of which I am chair) which
is looking at personalized versions depending on disability.

DavidF: Most latex is semantic, so goal is not to lose that semantics in
MathML

DavidF: pretex macro package helps authors in many ways. E.g., they avoid
having to change spacing before “dx” with “\,” inside of integrals along
with other issues in typesetting. Want to keep semantics implied by macro
name for accessible markup.

Daniel: there has been discussion in Aria WG of mapping between HTML tags
and ARIA, would we need something similar for mathml

Jason: if converting from tex or some other source, may not get a full
content mathml, so would you want to support partial semantics where you do
not have unambiguous content.

Charles: Sorry - recording not captured.

Didn’t get to the last item:

   - Indicating where insertion cursor or selection is (from Murray
   Sargent's blog
   post
   <https://blogs.msdn.microsoft.com/murrays/2017/03/19/editing-math-using-mathml-for-speech/>
   )