W3C home > Mailing lists > Public > public-digipub-ig@w3.org > October 2016

Re: [mediaqueries] MathML

From: Florian Rivoal <florian@rivoal.net>
Date: Wed, 5 Oct 2016 09:02:03 +0900
Cc: Paul Topping <pault@dessci.com>, "Tab Atkins Jr." <jackalmage@gmail.com>, Avneesh Singh <avneesh.sg@gmail.com>, www-style list <www-style@w3.org>, W3C Digital Publishing IG <public-digipub-ig@w3.org>, Peter Krautzberger <peter.krautzberger@mathjax.org>
Message-Id: <8951C554-4F03-48E4-B2C1-157F45B0A94A@rivoal.net>
To: Bill Kasdorf <bkasdorf@apexcovantage.com>

> On Oct 4, 2016, at 23:47, Bill Kasdorf <bkasdorf@apexcovantage.com> wrote:
> 
> I think we may be drifting away from the original intention of the MathML MQ request. So here's a recap.
> 
> --It came about because publishers, who often have MathML, do not put it in their EPUBs because they can't trust that it renders properly. They just put in images, which are inherently not accessible and typically not properly scaleable (they can be SVG but usually aren't).

I don't know how accurate that is, but the statement I heard was that while publishers who would like to distribute MathML as the main representation to end users are interested in this media query, publishers who have tried to do it were typically not happy with the result, and went back to using alternative methods, because neither rendering nor accessibility was reliably good.

> --MathML is _extremely_ widely used in STM publishing. It is considered the lingua franca for math. Perfect? No. But millions and millions of math equations currently exist, reasonably happily, as MathML within standard STM production and dissemination workflows.
> 
> --This means that it _does_ work for much of what it is used for; just not well in browsers and most EPUBs.

I am not disputing that MathML is tremendously useful. Only that a Media Query would make it more useful in browsers and EPUBs.

> --One thing it works reasonably well for is accessibility. I remember being told by a member of the accessibility community a couple of years ago that, in the case of a student needing a physics textbook in accessible form, the difference between the math being in MathML or just images can be a difference of $50,000 worth of work and six months of delay to get her that book in a form she can use. THIS MATTERS!!!!!

Yes, this matters. However, what I do not understand is how the media query would help:
 - You can already include the MathML in addition to the SVG/HTML/PNG rendering and make it visually invisible but reachable to the screen readers. In what way would a media query help you with that?

 - Peter has claimed (If I understand correctly) that MathML accessibility was worse than ARIA annotated HTML or SVG. Do you disagree? If not, why is that not the prefered approach.

> --We are looking for a way to make it safe for publishers to put their MathML in an EPUB so that it can be used for accessibility, even if it is not used for rendering visually.

Several people have claimed that they are already doing it today. What is the problem with the approach they use? Also, I really do not see how a media query would help there. If you have a visual rendering of the math done with something other than MathML, and MathML as well, and the MathML is styled to be visually hidden but reachable by screen readers, what would you do differently if you had that media query?

Keep in mind that a media query in no way helps you visually-hide-but-keep-accessible the content. It merely tells you, on the client side, about the device capabilities, so that CSS (or JS) can react to it. How would you react to it?

> --We are NOT looking to see if every possible aspect of MathML is correct. We just need to know whether or not it is there, and to provide the reading system with information about whether or not to prioritize using the MathML or using the companion image.
> 
> --MQ was suggested as a simple mechanism to enable that.

If you have both MathML and a companion image/SVG, you have both accessibility and reliable rendering. Why would you switch off the rendering of the companion image/SVG based on a browser's claim that it supports MathML? Since you already had a good rendering and accessibility, you gain nothing by doing so, and you risk making the visual part worse if the browser's MathML support has gaps (which is somewhat likely).

 - Florian
Received on Wednesday, 5 October 2016 00:02:32 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 25 April 2017 10:44:46 UTC