Minutes: MathML Core meeting 27 Jan, 2020

MathML Core Meeting of Jan 27, 2020



Attendees:

   -

   Fred Wang
   -

   David Carisle
   -

   Neil Soiffer
   -

   Patrick Ion
   -

   Murray Sargent
   -

   Rob Buis (15 minutes late due to conflicting meeting)


Regrets:

   -

   Brian Kardell
   -

   Bruce Miller


Agenda:
https://github.com/mathml-refresh/mathml/issues/8#issuecomment-576637211



The meeting was recorded:
https://zoom.us/rec/play/up0kfu6qr2g3TILBuASDCvMvW426Jqus2yAaqfMNxBzhVHFVNVSvNOZEauJUwBekAHaiBLOnV6VdX8N9?continueMode=true

Operator update?

No updates this week.

mglyph/malignmark & HTML5: #145 (comment)
<https://github.com/mathml-refresh/mathml/issues/145#issuecomment-576636538>
#25 <https://github.com/mathml-refresh/mathml/issues/25>

NS: we previously seemed to have agreed to keep mglyph in full but remove
from core

FW: but should we request that mglyph be removed from HTML 5

NS: there is a bigger issue dealing with polyfills

[ discussion of HTML and mglyph inside of an integration point and part of
parsing ]

FW: Also out of sync HTMLOrSVGElement VS HTMLOrForeignElement. So more work
needed to be done with specs there

https://github.com/whatwg/html/issues/4702

https://github.com/whatwg/html/pull/4893

NS: malignmark was probably never used (and barely implemented). A separate
discussion: should it be removed from full.

MS: it looks like we might use it in Office. I’ll look into it.

FW: issue with a polyfill for mglyph is getting the vertical align right.
(valign attr)

PI: I had some issues, but I’m willing to concede they are not important.

RB: MDN has good vertical-align overview
https://developer.mozilla.org/en-US/docs/Web/CSS/vertical-align

FW: but these are only for inline and inline-block and currently most
MathML elements are block.

PI: but the above-mentioned page explicitly says “Note that vertical-align
only applies to inline, inline-block and table-cell elements: you can't use
it to vertically align block-level elements
<https://developer.mozilla.org/en-US/docs/Web/HTML/Block-level_elements>.”

FW: maybe we could do some hacks using mpadded, either restarting the
MathML if there are other chars in the leaf or just remove the leaf if it
is the only character inside of the leaf.

FW: maybe better if in HTML namespace than in MathML so than just a regular
HTML element inside of a token element. Then could use shadow dom (after
being added to the allowed element list).

DC: useful in full because outside of HTML, there is no <img> element

Resolved: remove mglyph from core; stays in full
more on mrow-like: #183
<https://github.com/mathml-refresh/mathml/issues/183> #184
<https://github.com/mathml-refresh/mathml/issues/184>

FW: previously dealt with making some things space-like.

NS: your issues mentioned “unknown MathML elements”, but what about HTML
elements or unknown elements in general?

FW: Not sure. Spec says that if there is text inside something other than a
token element, it should be treated as if is wrapped inside of <mtext>.

DC: In HTML spec, it says non-MathML elements should be wrapped in merror  (
https://html.spec.whatwg.org/multipage/embedded-content-other.html#mathml)

FW: that still leaves the problem that it isn’t legal inside of merror :-)

FW: one option is that they are omitted from the render tree. Another
option is to layout using CSS.

RB: not sure what currently happens

NS: Off the top of my head, treating it like display:none is easier and the
way to go. Probably need to double check specs.

FW: Example a<math>x</math>b renders as ‘axb’ in Firefox; Webkit and Chrome
show ‘ab’.

NS: What if it were <math><p>x</p></math>?

DC: It breaks out of math:

https://www.software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cmath%3E%0A%20%20%3Cmi%3Ea%3C%2Fmi%3E%3Cp%3Eaa%3C%2Fp%3E%3Cmi%3Ex%3C%2Fmi%3E%0A%3C%2Fmath%3E%0A

NS: so this is already prescribed by HTML 5.

FW: but with JS, you can insert it:

    a<math><p>x</p></math>b

    a<math id="x"></math>b

    <script>document.getElementById("x").insertAdjacentHTML("beforeend",
"<p>x</p>");</script>

FW: Firefox will display it; Webkit/Chrome don’t.

RB: shouldn’t we always try to display the text?

NS: in semantics, we throw away all but the first child; this is somewhat
analogous in that we throw away anything but math.

FW: not really analogous because semantics has a defined semantics and can
use CSS to put display:none. Can’t do it with random elements.

FW: easiest to just not put it in render tree

Resolved: go with the Webkit/Chrome model and not display it.

FW: back to initial topic…

NS: handling math element is a little weird

FW: we can add an exception for that so that spacing is respected

NS: what about making mprescripts and none be space-like?

FW: has the problem that they would support height/width

To do: Read the issue and comment for next week.