W3C home > Mailing lists > Public > public-fx@w3.org > April to June 2012

Re: [Filter Effects][css3-transforms] Using MathML for formulas

From: Dirk Schulze <dschulze@adobe.com>
Date: Fri, 18 May 2012 13:13:44 -0700
To: Peter Krautzberger <pkrautzb@umich.edu>
CC: Vincent Hardy <vhardy@adobe.com>, Cameron McCormack <cam@mcc.id.au>, "public-fx@w3.org" <public-fx@w3.org>
Message-ID: <2C7A3D00-000F-4ABB-91B1-305CA19A727A@adobe.com>


Sent from my iPhone

On May 18, 2012, at 10:43 PM, "Peter Krautzberger" <pkrautzb@umich.edu<mailto:pkrautzb@umich.edu>> wrote:

Hello,

As part of MathJax, I would like to address Dirk's concerns. Short version: MathJax will only increase accessibility, not decrease it.

First off, accessibility is important to us and MathJax is designed with accesibility in mind (cf. http://www.mathjax.org/resources/articles-and-presentations/accessible-pages-with-mathjax/). We have dedicated features in this respect and in our experience screen readers have no problem with pages using MathJax. I should add, however, that we are not aware of screen readers that can read MathML without help from MathPlayer -- we'd be very interested to hear about such a project!
I am glade that accesibility is a priority. As far as I know, screen readers just access the DOM. Therefore if the content is not in the DOm, it can not be read by screen readers. I think you are right that there is no screen reader supporting MathML. Not necessarily a reason not thinking about the potential usage. Back to DOM. At the moment the code is not visible in the DOM, right? How can it be accessible then? I wonder if a alternate description could be helpful here with something like an alt attribute? Note that display: none in as style doesn't work either, since most screen reader ignore these tags. It would be great to here some input from screen reader implementers.


Regarding Dirk's concerns: MathJax does not remove the original MathML.when using MathJax's "HTML-CSS" output mechanism -- instead the MathML is heavily styled so that the browser displays it correctly. It is somewhat true that MathJax's SVG output mechanism removes the mathml from the DOM -- but it is added to the page in a script-tag after each SVG (and remains available internally for MathJax). In any case, we would not suggest SVG output as a default setting for the specs since it does not work with IE 8 and below.

That would also mean not to embed SVG in general. Not even for images. I think it is a mess to suggest authors to use SVG but we can't do it because a previous version of IE doesn't support SvG.

I would tentatively suggest to configure MathJax in such a way that "NativeMML" is used whenever possible (e.g. IE with mathplayer, Firefox if your code works there etc) and use "HTML-CSS" elsewhere; you can take a look at our documentation for specifying this http://www.mathjax.org/docs/2.0/options/MMLorHTML.html.


Yes, it should use native Implementations as much as possible.

Let me know if you need more information,
Peter Krautzberger.
Thanks for your feedback

Dirk


On Fri, May 18, 2012 at 11:05 AM, Vincent Hardy <vhardy@adobe.com<mailto:vhardy@adobe.com>> wrote:
Hello,

Is there a way to keep the work Cameron has done and address the accessibility concerns Dirk raises?

-v

On May 18, 2012, at 3:43 AM, Dirk Schulze wrote:

>
> On May 18, 2012, at 12:49 PM, Cameron McCormack wrote:
>
>> Without even having noticed this thread (Dirk just pointed me to it), I
>> added MathML supported by MathJax to the SVG 2 spec earlier today.  It
>> is currently referencing the CDN copy of the library (although the
>> stable version, not the "latest" link).  It works pretty well -- I've
>> currently got it using the SVG rendering mode, since it looks nicer in
>> than the HTML+CSS, IMO at least.
>>
>> I replaced one <pre>-formatted equation here:
>>
>>  https://svgwg.org/svg2-draft/painting.html#StrokeMiterlimitProperty

>>
>> Anyway, I'd love to keep using it in the spec.  I'd be fine with adding
>> a local copy of the library to the svg2 repo if necessary.
> (Previously posted on public-svg mailing list)
>
> I think it is to early to do this. For browsers that do not support MathML, the complete MathML code gets replaced by HTML or SVG as far as I can tell. Therefore it is no longer accessible by users of screen readers. That itself is in conflict with the idea of MathML and doesn't help at all. The semantics get lost completely. Not all people with the need of screen readers use Internet Explorer and the Math plugin. And there are still browsers like Chrome that don't support MathML. This becomes more problematic once we see laptops with ChromeOS and ChromeVOX. Therefore we should think about it twice. For the meantime I would like to see it removed from SVG 2 again untill we have a decision in the W3C with other WGs, including the MathML WG.
>
> Greetings,
> Dirk





Received on Friday, 18 May 2012 20:14:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 18 May 2012 20:14:14 GMT