Minutes: MathML intent meeting, 14 Jan, 2020

Attendees:

Neil Soiffer

Bruce Miller

David Carlisle

Deyan Ginev

David Farmer

Sam Dooley

Murray Sargent

Charles LaPierre

The meeting was recorded:
https://us02web.zoom.us/rec/share/tLcatjVCPn_HVROiYmD1RSENQteUOiFKcSVSgyKyiYx3xHVGUClG3c6wMAMgl2gp.tX0umhf8Pz7-01ne



Agenda:
1. Announcements/updates

MS:  I came up with a way to indicate the selection by adding some attrs.
This can be sent to a braille display.

NS: It could be fodder for the spec or for a note.

NS: Need new zoom link.

NS: EPUB WG wants to talk to us about concerns about the charter. Is
tomorrow at 10 Eastern ok?

<some yes>

CP: I think they just want to make sure nothing breaks

2. Deyan's action item on schema.org
<https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fschema.org%2F&data=04%7C01%7C%7C2d212f0fbdd941d0cb8208d8b6bea0a2%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C637460277147909164%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=FftoIbbWD188QKNtnuKzbpE1gfbl7sSKZ%2F4Zic67aM4%3D&reserved=0>

DG: I tried RDFa and JSON-ld. I was disappointed with the experiment. You
can request Google index pages. It takes less than a day to do that. See
https://dginev.github.io/tiny-mathml-a11y-demo/ for an experiment.

DG: If you do a site search, google finds the new pages.

DG: Their report on rich text support says everything was good on the pages.

DG: However, if you do a search for the metadata, google finds nothing. If
you search for something for something visible like h-bar, google will find
it.

DG: So it seems like they extract the data but don’t use it as part of
search.

DG: FYI: difference between the pages -- I tried using MathML and not
MathML.

NS: Disappointing for search, but it does narrow down the options.

3. Review of what we've looked at/Plans for future

NS: We went from mathrole to semantics to intent with a functional notation
using different variants

SD: As a summary. We had so many things for mathrole it would not make it
as an option to add to aria’s roles. Also, the role didn’t provide enough
info.

SD: … considered whether markup goes on the operator or to the outermost
level. [transpose example]

SD  … at this point, we worried about “semantics” as being too loaded in
meaning with everyone having their own thoughts about what it meant.
Transitioned to “intent”

SD … found places that just tagging the operators wasn’t enough. At this
point we moved to the functional notation with several shorthands. That
somewhat allowed minimal markup as with the earlier ideas of just marking
up the outermost part of the semantics

SD  … I think we got to the point where we can capture what we want to
capture

NS: DC had brought up using intent to deal with linebreaking and alignment
done with tables.

BM: The mystery to me is that we have RDFa-like semantics. We need
structure for computation and maybe for speech. That remains unclear to me.
The intent proposal is heavyweight, so it’s not clear that’s what you want?
Why not just use content MathML if that’s what you need?

DG: Two things I wanted to get clarity on. One is scope and the other is
applications. I’m still a proponent on open-endedness. The split into level
1 and everything else where level 1 is common and known and let’s make sure
that is designed/described so everyone is on the same page.

DG: … I’ve been working on level 3 so people who want to deal with more
advanced math or other topics.

DG: … On the computation side, I wanted to get clarification and it is
becoming clear. I’m certain that we now need to prioritize accessibility
over computation. The intersection of search and computation and
accessibility might be fragile.

MS: I agree with the importance of accessibility

NS: For accessibility, I don’t need all the structure but it doesn’t hurt.

DF: Accessibility is why I’m involved. I understand computation is
interesting but I’m not sure how that will work. My goal with PreTeXt is to
make sure we don’t lose.

BM: I have a tentative request. It would be informative to me to get a
sense of what information is needed for accessibility. I don’t want to get
bogged down into the details on how to do the markup, just a higher level
need.

NS: I’m happy to do that. What would you like to have annotated?

SD: so what you’re looking for is what is the minimal info over
presentation MathML that is needed for a11y?

BM: Yes, I don’t think we need to reinvent content MathML. We are looking
for something simpler, but what.

DG: I like the discussion of what we need for speech.