Minutes: 18/11/19 core meeting

MathML Core Meeting of Nov 18, 2019



Attendees:

   -

   Fred Wang
   -

   Neil Soiffer
   -

   David Carisle
   -

   Rob Buis
   -

   Patrick Ion
   -

   Bruce Miller
   -

   Brian Kardell


Agenda:
https://benetech.zoom.us/recording/share/7DA8siVRCggwTafv6yp4AaBI1H-jtpfGwQBSBW5bzgGwIumekTziMw


The meeting was recorded:
https://benetech.zoom.us/recording/play/bRdNIKVYjVcFVNG2GQz0eU1AacqmfCzOb71k4484vHoFC7xTPxujAQSD4MFJBXuW?continueMode=true


Quick updates

MathML 2020 Conference / Hackfest
FW: Maybe organize conference at the same time/venue as the Hackfest (May,
2020). Nothing definite.
[a few people on the call said they might come to such a meeting]

Chromium

RB: started upstreaming. Still working out the details of what goes in
first. Lots of reviews. Need to put in smaller patches. Also noticed things
to improve on the internal branch.

RB: Have some ideas who will be doing the reviewing. Gotten good feedback.

NS: The fact that you have reviewers doing the reviews is a great sign for
interest.

RB: Hard to do in small patches and keep the patches simple. Probably will
be a little stop and go as I learn about what is acceptable and what
changes the reviewers want.

BlinkOn 11 - two MathML talks (brian, fred) will be livestreamed/recorded

FW: Display math suggestion change
<https://github.com/mathml-refresh/mathml/issues/171>. A new inline math
value apparently comes along for free. I’ll leave it to BK to comment
further.

FW: Positive feedback from Ian Kilpatrick (Layout TL) and Chris Harrelson
on MathML from the chrome layout team
<https://www.chromium.org/teams/rendering#TOC-Team-Members>

FW: Also had a meeting with someone on the a11y google team (Dominic
Mazzoni). Happy with idea of parity with webkit and gecko.

FW: We had two presentations. Mine was similar to what was in the hackfest.
Updated a little. Positive feedback.

NS: this is better to share than the hackfest version?

FW: Yes. Talk is at https://www.slideshare.net/igalia/. Positive feedback
on twitter <https://twitter.com/briankardell/status/1195069556684775424>
(and this one <https://twitter.com/briankardell/status/1195063985512411137>
).

FW: BK did his talk during the flash talk -- 35 slides in 3 minutes(!)

BK: Talk is here
<https://docs.google.com/presentation/d/1Kg-IEWUJqD4fj71EkYxTtM7QKiHkaim6xgO5MIEY7Qs/edit>,
also positive feedback on twitter
<https://twitter.com/briankardell/status/1195530559901929475>

TAG Review process w3ctag/design-reviews#438
<https://github.com/w3ctag/design-reviews/issues/438>

BK: It is on the agenda. We have reviewers assigned: Dan Applequist, Hadley
Beeman and Alice Boxhall

BK: talked with some TAG/former TAG at blinkon/dev summit - generally
positive feedback.

------------------------------

Clarification for Operator maxsize/minsize: #110
    <https://github.com/mathml-refresh/mathml/issues/110>Keep current
definition?

NS: I added a comment that says MathPlayer does scaling. But there is an
issue of rounding because of the division. Does rounding come up elsewhere
in layout?

FW: This is probably the only case where rounding matters.

BM: Does either mess up symmetry?

NS: Both methods should be ok.

All: some clarification of what is meant.

Resolution: do  (minsize / (height + depth) , and then calculate depth as
the remaining space for the target size.


Operator spacing in scripts (38
<https://github.com/mathml-refresh/mathml/issues/38>)
    Proposal 11: Keep current definition? (Neil)

DC: I asked a few TeX people whether not following TeX was ok.
Surprisingly, no negative feedback. Not likely to get pushback from TeX.

BM: In the examples, it looks like the TeX renderer used some spacing,
which I like.

BM: Maybe the fonts contain a little spacing around the “+” etc?

PK: The important thing is to be able to read it

BM: I’m in favor of the Gecko rendering as it guarantees some spacing if
there is no space around the character in the font.

FM: From an implementation point of view, the Gecko solution is pretty easy.

NS: DC thought maybe they could use CSS if they want no spacing (set
lspace/rspace to 0 in scripts).

DC: [responding to BM’s earlier comment] I looked at TeX, in classic TeX
the script font (cmr7) is noticeably wider than 10pt scaled 70%, which
means that the effects of TeX not adding space are less pronounced. With
xetex/luatex and a scaled font (Stix Two here) it is perhaps more
noticeable and having the renderer add some space might be better. Here are
details:

Cmr


Stix2


Images made with pdflatex and xelatex from

\documentclass{article}

\usepackage{graphics}

\ifx\Umathchardef\undefined\else

 \usepackage{unicode-math}

 \setmathfont{STIX2Math.otf}

\fi

\showoutput

\begin{document}

\setbox0\hbox{$\scriptstyle +$}

\setbox2\hbox{\scalebox{.7}{$+$}}

\showthe\wd0

\showthe\wd2

\[ 1+2 \quad  x^{1+2} \]

\end{document}

PI: TeX comes with a system of fonts where they are compatible.

DC: Maybe LuaTeX looks tight with STIX fonts.

FW: Let me experiment with some other systems. (Gecko, Chromium,
LuaTeX/XeTeX with latin modern math)

NS: postpone to get more examples

Operator Dictionary update:

FW: currently about 1100 entries. Naive implementation is binary search.
Google person recommended try a perfect hash.

FW: need to make some of values more consistent. Probably some errors.
After that, maybe we can write a more compact form.
https://github.com/mathml-refresh/mathml/issues/176

DC: we should make it more compact. One interesting thing is continuous
runs. Maybe we could include more code points.

NS: We should not include the table in the spec. We should point to the
data as is done in the Unicode specs

FW: HTML includes a table for character names

DC: I don’t think there is a problem including a JSON form.

DC: I’ll try and make some proposal on changing the table

FW: At least we have a proposal for the Arabic chars

FW: Probably true for accents

NS: Some group of people needs to go through the table and clean it up.

FW: I reorganized and the outliers might be more obvious.

DC: The table was put together from lots of sources and that leads to
inconsistencies. It was hard to see the errors with the previous formats.

DC: Integrals seems to be two different groupings, and that doesn’t make
sense.

NS: Anybody what to volunteer to try and do some cleanup?

DC: I’ll take it on.

Meeting next week.