Minutes: MathML intent meeting 15 Oct, 2020

The meeting was recorded:
https://benetech.zoom.us/rec/share/5AcuVKSD-X4ft145XS2AxX1OC_v4MgKym8bjp9oncQD1ZW_g7K9592m8zMqBlYTR.W9fUJD4N9MYIT4Sj
Passcode: 4b&%WcyP  (starts about 8 minutes in)


Attendees:

Neil Soiffer

Deyan Ginev

Sam Dooley

Patrick Ion

David Carlisle

Murray Sargent

David Farmer

Steve Noble

Bruce Miller

Patrick IonN

Regrets: Louis Maher

Thanks to a mystery note taker for helping out...


1. Charter (https://mathml-refresh.github.io/charter-drafts/math-2020.html)  a)
Take a vote on Moritz's concerns for the charter per the discussion at the
last meeting (
https://lists.w3.org/Archives/Public/public-mathml4/2020Oct/0003.html)

DC: We can’t really add any other than attrs. But I don’t object to
generalizing more.

NS: I agree we take out mention of a “intent” attr

DG: RDFa was discussed as an alternative.

NS: That’s sort of using attrs.

DC: Let’s just delete “add attributes that” in the MathML 4 part.

NS: Also in the intent part of to do items

Consensus: agreed to do the above
  b) Other charter comments?

DC: If we want to make a PR back to th W3C chater draft repos
https://github.com/w3c/charter-drafts/compare/gh-pages...mathml-refresh:gh-pages

NS: I’ll check with Bert.
2. Continued discussion on "intent" syntax/usage

Action Item: NS to reach out to Moritz to see if he want to present an RDFa
proposal to the group.
   a) continue discussion of David C's idea to deal with alignment

Neil, most people use tables to get alignment (rather than maligngroup)

DC had suggested that could use intent= on rows to “reconstruct” the
meaning of the expression.

Sam: Interesting possibility but we would need some way of merging the
intents on individual rows.

Neil mostly it is concatenating the columns together and then somehow
merging the rows.

Neil/Bruce eg system of equations.

Deyan eg piecewise function

Bruce you need to encode the meaning, the mini-language we were suggesting
specifies how to assemble the pieces

Neil: action to Deyan to add a piecewise function and system of equations
using a table but with the extra markup for intent

Sam: matrices with elisions another example.

Deyan: Elisions are not math constructs but linguistic devices so they have
not appeared in lists generated so far.

NS: Elision happens in series, etc., outside of tables

BM: It isn’t just removing simple mathematical terms. Sometimes operators
are part of the elision.

DC: 1, 2, …, 5 as opposed to 1,2,.....   Where the trailing … is not referring
to an omitted block of text

DG: Identity matrix is another example. Probably don’t want to hear all the
details once you know what an identity matrix is.

   b) review/remember where we are at on intent syntax proposals

NS: I think we were moving toward Bruce and Deyan’s proposal but Sam had
some objections.

SD: I agree but I want to have something that says go find the pieces.
Making it simpler. E.g, if we say it is an integral, but we don’t use the
functional notation. The children are still tagged.

BM: Maybe we have transpose(...) and some other way of implicitly saying
“transpose”

SD: Name the main operator, and say the subexpression is an argument of its
parent. Maybe numbered arguments vs named arguments.

NS: What’s the advantage?

SD: Most cases happen by default.

SD: Agree that it doesn’t solve all the problems.

SD: Consider int x^2 dx. Mrow for integral would be intent=”integral”. And
mark the x^2 it as intent=power and mark the x as intent=variable. The
parts marked with “intent” would be automatically collected together as the
arguments of the integral.

NS: Actually we talked about intent and arg, where ‘arg’ could be ‘1’, ‘2.

NS: Doesn’t solve the case with 1/r as the integrand.

SD: Yes, I want to two cases -- a common way to make it easier for most
cases.

BM: That would make the harder cases harder.

Action Item: Sam will update his examples and send them out again.

We meet next week