Minutes (corrected): MathML intent meeting, 29 Oct, 2020

The link to the recording was wrong in my previous email. Here are the
minutes with the correct link...


The meeting was recorded:
https://benetech.zoom.us/rec/share/BNDOH-4S54OcN9krLEA5A2g1ylS9ZwFUoFylPJY8l9ZfK4tMEvjHN0-p4y0Gws4J.25OwSumCUehZrpNj
Passcode: A#y+e9Rw (starts about 15 minutes in)


Thanks to David Carlisle for doing the minutes.

Attendees:

Neil Soiffer

Deyan Ginev

Sam Dooley

Patrick Ion

David Carlisle

Murray Sargent

David Farmer

Steve Noble

Bruce Miller

Patrick IonN

Louis Maher

.

Summary of TPAC Call

NS: Thanks to DC for minuting the breakout session on IRC!

Minutes https://www.w3.org/2020/10/29-math-minutes.html]

NS: We had 45 attendees at the breakout session, which was spectacular!

NS: Brian gave a presentation on status of MathML

NS: I spoke about what is in mathml core and what is in full, and about
charter.

NS: discussion mentioned apple not wanting to drop things due to non web
uses, some questions about links and about semantics.

NS: one person (Florian) concerned about English bias in the table with
semantics.

NS: at end of meeting Brian asked for sense of meeting whether people were
in favour has around 12 +1 and no negative comments

Deyan: nice to see Ian K. at google commenting favourably on current igalia
work that it looks like being maintainable and not like previous attempts.


Sam’s Proposal

I posted a link to https://mathml-refresh.github.io/mathml/docs/intent.html

SD: discussion of rules for distinguishing function from expression of
function applied to no argument

BM: the rules were not clear from the document

NS: Thing that confused me was an msup with intent power but power has no
arguments

NS: how do you distinguish function application.

SD intent = name refers to the function, as opposed to intent=power which
would be function application to the children which have their intent.

SD: this makes the common cases simpler to mark up.

MS: simplicity is key it is hard to imagine implementing all the argument
syntax,

BM: it is easier to markup if you don’t have to add intent in some places
but it shifts the problem to the consumer. I can not tell when to look for
children and when not.

MS: for accessibility you don’t need full markup eg “ackermann’s function,”
you would just say A

BM: but it isn’t clear if we mean just A or A of ..

SD: what makes it clear to me is that you can always use the arguments in
the intent. If you use the short form intent then if it has children then
it applies function application

MS: we can use mo with U+2061 to denote function application. Similarly use
U+2147 for exponential e.

DG: we have examples but we don’t have a list of rules. It would be good to
write those down then we can check that all cases are covered.

NS: useful to try an implementation at some point as that clarifies the
special cases.

SD: I hoped the document example helps tease out details of dealing with
names of functions as opposed to function application,etc.

BM: in the power example the children don’t have arg or intent, if you have
msup intent=foo, is that foo of  x and n, how would you right the msup is
simply foo

NS: how do you mark up that the entire msup is a notation for “foo”

SD: You can write out the notation without intent on sub-expressions. Could
have possibilities like intent=none to remove the implied intent on mi.
(not in current document)

DG list of mesons in the table where the whole subsup is the named construct

BM: basically this corresponds to “embellished symbols”

NS: in chemistry they have decorated arrows with similar.

SD: in the document I assume mi infer a variable and mn defaults to number,
could possibly not have these defaults or make the defaults depend on the
subject area.

DG: we need the whole list including defaults early to allow discussion.

NS: we can talk about this in two weeks, Need to see Deyan’s presentation.

DG: I think we should have defaults that make it easy to get 90% of the way
there but we need to be careful with the definition.


Deyan’s Work

DG: Examples of alignment in shared screen, with spoken renderings (
https://dginev.github.io/tiny-mathml-a11y-demo/)

DG 11 component expression in the annotation on the outer mtable, but
children are in different levels in the presentation

DG: the top level alt annotation gets very complicated

MS: it is sufficiently complicated that it won't get mainstream
implementations

DG: show speech with and without semantic annotation.

DG: discussion of difficulties of inferring  if an operation is infix
R(a,b) or a R b when teh structure is at different levels in the mtable

MS: OfficeMath writes for the two equations x+y=z and a=b+c aligned at the
equal signs

<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="block">

<mml:mtable>

<mml:mtr>

<mml:mtd>

<mml:maligngroup/>

<mml:mi>x</mml:mi>

<mml:mo>+</mml:mo>

<mml:mi>y</mml:mi>

<mml:malignmark/>

<mml:mo>=</mml:mo>

<mml:mi>z</mml:mi>

</mml:mtd>

</mml:mtr>

<mml:mtr>

<mml:mtd>

<mml:maligngroup/>

<mml:mi>a</mml:mi>

<mml:malignmark/>

<mml:mo>=</mml:mo>

<mml:mi>b</mml:mi>

<mml:mo>+</mml:mo>

<mml:mi>c</mml:mi>

</mml:mtd>

</mml:mtr>

</mml:mtable>

</mml:math>

This seems to be simpler than using a more detailed table.