Minutes: MathML Core Meeting, 6 July, 2020

Attendees:

   -

   David Carlisle
   -

   Neil Soiffer
   -

   Brian Kardell
   -

   Murray Sargent
   -

   Rob Buis
   -

   Bruce Miller
   -

   Louis Maher
   -

   Patrick Ion
   -

   Frédéric Wang



Agenda:
https://github.com/mathml-refresh/mathml/issues/8#issuecomment-653411195

The meeting was recorded:
https://benetech.zoom.us/rec/share/3-NZN6325z5IG7PpwVvBZpY8GavLT6a8hCIa_PULz0-eN9mzY__FhVmX6bpc4fAT
Password: 0g$T+0q2

We ended up spending the entire meeting talking about links (again).
href (#125 <https://github.com/mathml-refresh/mathml/issues/125>) [TAG
feedback?]

BK: The preference of the TAG was to punt on links. They felt that the
safest move is to not have any links in MathML and instead have a polyfill
use JS and events. Maciej from Apple would prefer to see a dedicated link
element.

NS: won’t this break existing content?

BK: Firefox and Safari won’t stop shipping support, so it won’t break
current support. It won’t be interoperable.

DC: This is bad. Why can’t we have a new element?

BK: <a> in MathML works now, but display is messed up. See
https://codepen.io/bkardell/pen/NJjJMR?editors=1000

BK: lots of limitations on it. Doesn’t support ‘target’.

BK: too much to add a <ma> in core, deprecate href on all href

BK can put a link around the MathML or stick it inside a token.

DC: but that’s not useful.

DC: that essentially removes links from MathML because MathML is converted
to non-web contexts.

BK: I don’t think we can get <ma> into the spec

DC: but at least we could at least polyfill it. Removing all links removes
any way to have links.

BK: we can’t get it into all the browsers this year. I think a JS based
event is the only viable solution.

FW: we say unknown elements as mrows, so they would get the ma element for
free since all elements allow href.

NS: other browsers don’t support that yet

FW: but it’s not controversial, href support is

DC: <ma> would be more interoperable

BK: everyone wants that, but can we add that for level one

BK: I think that is the route that we should pursue. Can we get it done?

BM: I’m concerned about a new element, so core becomes a superset of a
subset. That breaks existing MathML code outside of the web.

FM: why don’t semantics, annotation-xml have a “m” prefix.

DC: they weren’t part of presentation MathML. Content doesn’t use ‘m’

BM: I’d rather leave things as they are. Chrome would look a little shabby
in comparison to Firefox or Safari.

DC: Don’t really understand why it isn’t acceptable to have an href but it
is fine to have an onclick event.

BK: If we said we would like as a goal to have a link element as the way to
do it, does this small group agree to it?

DC: it’s not my preferred solution.

NS: I agree because of compatibility issues.

PI: If that is only solution

MS: I’m game for any solution

FW: My preference is to a have link element in core, but not level 1

BK: add some prose in core spec that this is what we would like to do and
this is how you polyfill it.

NS: I’m dubious anyone would change something for a thing that doesn’t work.

BK: that’s what people thought of at the end of the tag call.

BK: I’m really game to get a real link element.

DC: It’s not the best solution for MathML use though.

BK: can people live it

DC: I can

BM: I probably wouldn’t use it; I’d probably keep using href and polyfill
if needed for Chrome.

BK: I think we are dealing with levels of “ok”ness. Implementers want to be
as conservative as possible..

BK: I think we will end up with ‘a’ plus token elements.

BM and DC: that’s better

BM: If you introduce ‘a’, does that take away incentive to support href on
tokens?

BK: Maybe a little. But the ecosystem is going to dictate this.

BM: I don’t they are mixable. href on element is the most important.

BM: Where are we on href on tokens?

BK: The issue is can we get it implemented in Chrome? Can we get it
implemented in other browsers?

BK: I think the CG agrees that we don’t want href on mrow because we want
mrow to be extensible, so we need <ma>, but we want it on tokens.

BK: table cells are a bit of the same thing -- you can’t put a link on td.

FW: can we put in core that href is supported on token elements?

DC: that would be better; not ideal but I can’t get everything.

BM: it would be good for me

DC: while we all want links, we want core to move forward in both spec and
implementation.

FW: we either add it to token elements or polyfill it in the short term

NS,DC,BM: does it really need to handle all the link attrs?

BK: I think google wants it all

BM: whatever is strategically best is what we should pursue?

FM: since not all links attrs are supported, this might be hard to get done.

BK: TAG wasn’t dead set against it being on token elements.

BM: It is a shame if we lose this because we can’t get the bells and
whistles to work (the non-href stuff).


*Resolution*: we will try to do token elements with just href in core level
1; ma/a might be level 2 (see
https://mathml-refresh.github.io/mathml-core/#ref-for-dfn-mathml-hyperlink-elements-1
)

BK: if for whatever reason we can’t get it done by (say) Jan 1, do we still
want to mention it?

FM: we can put it in list it as “at risk”