Re: Agenda for MathML Core meeting 2019/03/18

Hi everybody,

Thank you for yesterday's meeting. I think we resolved most of the
issues here:

On 13/03/2019 20:17, Frédéric Wang wrote:
> (1) scriptsizemultiplier and scriptminsize attributes: Is it really
> useful for users to specify them or in practice or can default/font
> values be used instead?
>   See
> https://mathml-refresh.github.io/mathml-css-proposals/math-script-level-and-math-style-comments.html#scriptsizemultiplier-and-scriptminsize-attributes
>   PROBLEM: In the former case, that would mean implementing more CSS
> properties.

As I understand, we can probably remove these attributes from MathML
Core:
https://github.com/mathml-refresh/mathml/issues/1#issuecomment-474261094

> (2) Is is really important to have a script min size in MathML?
> Alternatively we could either (a) set a max value for script level
> (decrease is exponential, so it should be small like in TeX or Microsoft
> Word) or (b) continue to decrease font-size and rely on other mechanisms
> to address the problem (UI config?, a11y?)
>   See
> https://mathml-refresh.github.io/mathml-css-proposals/math-script-level-and-math-style-comments.html#scriptsizemultiplier-and-scriptminsize-attributes
>   PROBLEM: In addition to the actual font-size (constrained by script
> min size), CSS code has to track the unconstrained one otherwise we
> can't go back to the original size by increasing the script level.

People agree that the use cases of script min size (accessibility,
readability) are important. However they are addressed by the new
https://drafts.csswg.org/css-fonts-4/#propdef-font-min-size which has
better CSS definition and interaction with the rest of the CSS code. So
we could rely on this property and defer the details to the CSS WG.

> (3) Define exactly how the font-size/scriptlevel interacts in term of
> CSS inheritance. Paragraphs from [2] [3] indicates the font-size change
> when font-size is not specified. When font-size is specified, MathML
> full [3] seems to say we should apply script level before font-size
> change. I personally think order is not so important: (a) Changing
> font-size *inside* a math formula is probably not super useful in
> practice (b) In any case, people can decide to wrap the font-size change
> in an mrow-like element (e.g. non-browsers would do <mstyle
> mathsize="...">) avoiding the interaction with scriptlevel.
>   PROBLEM: The choice of interaction can make code complicated [1]. We
> should probably choose the most simple option here.

It seems what Firefox does after Emilio's latest change (
https://bugzilla.mozilla.org/show_bug.cgi?id=1534494 ) is what the
MathML CG wants, basically font-size (mahsize) overrides
script-math-level changes (scriptlevel). I would try to align the CSS
scriptlevel proposal on Firefox

> (4) Script level increment in munder, mover, munderover scripts
>   See
> https://mathml-refresh.github.io/mathml-css-proposals/math-script-level-and-math-style-comments.html#script-level-increment-in-munder-mover-munderover-scripts
>   PROBLEM: Scriptlevel change cannot be implemented with pure CSS, so
> this is adding some hacks to Firefox, violating layout code invariants [4]

This is undecided. On one hand it would be a serious MathML
change/polyfill but on the other hand it is a serious CSS/layout violation.

> (5) Should we take OpenType MATH values into account?
>   See
> https://mathml-refresh.github.io/mathml-css-proposals/math-script-level-and-math-style-comments.html#opentype-math-values-scriptpercentscaledown-and-scriptscriptpercentscaledown
>   PROBLEM: It is unknown whether the CSS code can read the font tables.

It seems the CSS engine has access to font data, so we can could use
them (and hence replace scriptsizemultiplier's use case).

-- 
Frédéric Wang

Received on Tuesday, 19 March 2019 09:35:17 UTC