MathML core meeting, March 18 notes

The meeting was recorded and can be heard at
https://benetech.zoom.us/recording/share/OS3qrX_UZ6PwT6n4Z2pDH9Bs1VU5phMV2DAcL1G0cGGwIumekTziMw

Meeting summary:

The initial part of the call went over the deprecation issue listed in [1]
Remove thin thick medium (mfrac)
David: take out of core, leave in full
Neil: there are some precedence for named values (eg in css) but
agree, remove from core.

Remove small normal big (mathsize)
David: as above
Neil: we could note in main spec that non core features will require
polyfills so you may not want use
general agreement.

Remove non 0 unitless values
gone from core
deprecate from full

Remove "other" attr (already deprecated)
remove "other" attribute, support data-zzz attributes (valid but browsers
will ignore by default)

Remove "macros", "mode" attributes
Remove from spec (already deprecated)

mglyph
#5 deprecated attributes fontfamily (drop)

token elements
deprecated attributes fontfamily fontsize etc


named attributes thinmathspace
remove from full (and core)

as a general principle we are removing named lengths suggest remove
these
action: open gh issue on named lengths.

#55 obtain usage stats on attributes (still open)

menclose radical
drop from core agreed.
proposal: deprecate in full (no disagreement)


mglyph
keep in full for non web contexts.
action: continue to discuss in #25
action: try to write polyfill and see what issues are with dropping it from
core.

#26 maction
Murray: uses it for current insertion point and showing selection points.
Neil: you could use attrs on token elements, mrows, etc., instead
No resolution



#31 (and email from FW)
scriptlevel, minsize,  scriptsizemultiplier causes some problems
TeX and MATH font table only specify two levels of shrinkage; MathML allows
arbitrary shrinkage
MATH font table effectively specifies scriptsizemultiplier
minsize interacts with the minsize specified in browsers
current size changes depending on font and language
discussion on interation with css or browser min font size
Fred: have gh css proposals, some people in css suggested to move to css wg
tracker.
current version doesn't mention minsize.
Emilio: currently scriptlevel handled mostly in the user-agent stylesheet,
Emilio gecko fontsize over-rides scriptlevel settings.
Neil: (checking spec) -- that's the right thing to do
Fred: CSS reads some font values such as 'ex' and 'em'. Can CSS read the
values (eg script size multipliers) from the MATH font table?
Emilio: should be possible


Fred: there is a problem with the CSS rules for changing scriptlevel for
mover, munder, munderover when the 2nd/3rd child has accent=true. CSS rules
can't access the content to determine whether the character should have
accent=true.
Bruce: can the operator dictionary be pushed into CSS
Fred: wouldn't help.
Fred: would like to remove from core, could use a polyfill that sets the
attribute
Neil: that would mean a lot of math would need a polyfill because of things
like "x hat". We would need to get sites to always use accent=true.
No resolution


[1] https://github.com/mathml-refresh/mathml/issues/8