WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: MathML and simple maths content

for

From: jeffgutsell@fuse.net
Date: Jun 2, 2019 9:33AM

The sorry state of math support is a good topic for an article at webaim.org. At a minimum, authors should be warned of the challenges. I did some searching on the Web and was struck by the skimpy resources. Even universities seem to have old, incomplete guidance for their content developers. For example, I found an MIT page that refers to HTML 4.1 being in draft status.
Jeff
Gutsell
.


-----Original Message-----
From: WebAIM-Forum < <EMAIL REMOVED> > On Behalf Of Mallory
Sent: Saturday, June 1, 2019 4:14 AM
To: Isabel Holdsworth < <EMAIL REMOVED> >; WebAIM Discussion List < <EMAIL REMOVED> >
Subject: Re: [WebAIM] MathML and simple maths content

Hi,

When I was working in education, while we tried to do everything possible in MathML, what most simpler math students wanted and needed was the so-called ASCII math. Similar to what you'd type. We did this with MathJax, which was pretty heavy for phone CPUs. Iphones were not happy with long pages of equations (which is why everyone would love not to need MathJax, including the creators).

It's this ASCII math that our alt text was made out of most of the time. It's good to have an instruction telling students to up their verbosity, so they are more likely to hear something like
−4 + −2 = −6

With lower verbosity levels, many symbols get ignored.

Our alt text switched to words once you got to phrases like "square root of".

As someone mentioned earlier, the alt text is a firehose. It cannot be inspected by the student bit by bit, so alt text was not recommended for larger in-depth formulas unless the formula itself was being presented in bits (which we had a lot of, showing how to reduce fractions or simple algebra steps).

For things like PowerPoint (which are unfortunately very popular), the creators needed to use MathType to render the equations, while students were required to purchase MathType for $99 (special academic price).
Normally, for web, the free MathPlayer was enough, but for MathML to be rendered in something using OLE's like PowerPoint, the student needed the full MathType (and not Lite, which was the fallback after the 30-day free trial). Or, this was the state of things a few years ago-- DS/Wiris may have been working on making the Lite support MathML in OLEs after version 4.

Btw, this is now WIRIS, they've bought or teamed up withDesignScience, makers of MathPlayer (which btw worked pretty well with
NVDA+Firefox for many years. Chrome never worked as at the time it had
zero support for MathML. For Edge, MathML support is a top voted request request on UserVoice, a place where people can vote for features:
https://wpdev.uservoice.com/forums/257854-microsoft-edge-developer/suggestions/6508572-mathml).

You'll need a combination of MathML and a kind of alt text/ASCII math for math under Algebra 1 or 2... once you get around that level or higher, MathML breaks down. Technically, the spec could probably handle it, but apparently it gets very complicated very fast and even browsers who want to support it run into trouble quickly. MathJax was still being used heavily but above certain math levels it too started having bugs displaying or describing more complex formulas. Quadratic formulas kind of things.

For higher math, today it seems the only thing that always works all the time, allowing not only students to read math but also to write and format it so that teachers and fellow students can read it as well, is LaTeX. It always works, but it's verbose, and requires both student and teacher to know it.

For Braille support, Sam Dooley did make an online math palette thing which can switch between displaying rendered math (via content MathML, which then presented on the web page as presentational MathML) and Braille.
http://accessibility.pearson.com/aee/
I'm not sure what the status of this is right now, and of course it's only for the issue of Braille, but it might be nice to know it exists if you're dealing with Braille users and doing that 3-day turnaround thing where someone prints it, someone reads it, results sent back thing.

I had no idea Brian was working at Igalia or that Igalia was working with the chromium team. This is fantastic news. I keep hearing rumours of the death of MathML due to the lack of support in browsers. I think the SR vendors can only really correctly read math once browsers are all exposing it in generally the same way. But I don't know the solution for when math starts getting difficult/complex and the world stops agreeing on how certain things should be written or spoken out loud. Most students say they'd prefer they hear math the same way teachers and classmates speak it, but there are global and even regional differences (I don't know the details of this, I've heard this from blind math whizzes).

cheers,
Mallory

On Thu, May 30, 2019, at 12:02 PM, Isabel Holdsworth wrote:
> Thanks Brian. This sounds reasonably encouraging.
>
> On 29/05/2019, Brian Kardell < <EMAIL REMOVED> > wrote:
> > On Wed, May 29, 2019 at 9:29 AM Birkir R. Gunnarsson <
> > <EMAIL REMOVED> > wrote:
> >
> >> MathML support is spotty/inconsistent and sometimes downright missing.
> >> Kepe inmind that IE11 is eventually going to g away, probably in a
> >> matter of years, so we, as a community must focus on
> >> Chrome/Firefox/Edge (which will run on Chomium).
> >>
> >
> > I'd like to add to this that mathml has a complex history and
> > suffered a series of very unfortunate events which I wrote about on
> > my blog https://bkardell.com/blog/Math.html, if you are interested
> > in the context around this, including the current state an efforts.
> >
> > To sum up a TL;DR of the current state:
> >
> > My organization, Igalia, contributes (in the code sense) to all of
> > the browsers and has been working with Chrome developers for a few
> > years to work out some kind of proposal. There was some general
> > agreement of a plan. We began a fundraising campaign
> > https://mathml.igalia.com/. We received initial funding from the
> > NISO through a grant from the Alfred P Sloan Foundation. This plan
> > was sent for review to the W3C's Technical Architecture Group
> > (https://github.com/w3ctag/design-reviews/issues/313),
> > and feedback seems positive. There is an active community group
> > doing all of the necessary spec work and hard decisions involved in
> > the plan https://mathml-refresh.github.io/ . Part of this work is
> > to ensure interoperability between implementations as well. Our
> > implementation of this work in Chrome is making good progress as you
> > can see here
> > https://mathml.igalia.com/news/2019/05/03/igalia-is-making-good-prog
> > ress/#new, and we would love for more people and organizations to be
> > involved. If you, or your organization thinks this is important
> > work and would like to chat about how you can help, feel free to
> > reach out.
> >
> >
> >
> > --
> > Brian Kardell :: @briankardell :: bkardell.com
> > > > > > archives at http://webaim.org/discussion/archives
> > > >
> > > archives at http://webaim.org/discussion/archives
> >