# [MathML] Re: The AAP EPUB 3 Implementation Project

From: Ivan Herman <ivan@w3.org>
Date: Thu, 7 Nov 2013 07:08:51 +0800
Cc: Richard Schwerdtfeger <schwer@us.ibm.com>, Gerardo Capiel <gerardoc@benetech.org>, Peter Krautzberger <peter.krautzberger@mathjax.org>, W3C Digital Publishing IG <public-digipub-ig@w3.org>, W3C Public Digital Publishing IG Mailing List <public-digipub-ig-comment@w3.org>, Bert Bos <bert@w3.org>
Message-Id: <1BF4E6D7-4234-4278-AEF5-7EE74B1035F4@w3.org>
To: Jean Kaplansky <Jean.Kaplansky@aptaracorp.com>

On Nov 6, 2013, at 23:39 , Jean Kaplansky <Jean.Kaplansky@aptaracorp.com> wrote:

> Dear Digital Publishing Interest Group -
>
> Are these on-the-go math conversations something you want to see continue on this list, or should we take this off line and then come back to the group with use cases?
>
> I ask because I think we can keep the conversation about MathML, MathJax, and Accessibility going in great detail, but I certainly don’t want to hijack the W3 DPUB threads for this purpose if we should take the conversation off line.
>
> Now… If MathML is not your hot interest, you can stop reading this message here. Please read on if MathML and the related issues, use cases, solutions, etc., is a big area of interest to you…

MathML was, many years ago, the first W3C Spec that I thoroughly learned, before even joining W3C. So...:-)

>
> Please get in touch with me individually at jean.kaplansky@aptaracorp.com?
>
> I want to start a list of people who want to continue this conversation and then talk to Ivan and the W3 powers that be about potentially starting an official MathML sub group (beyond just the STM group) as part of the Digital Publishing Interest Group. Also, if you know of people who should be on this list but who are not currently part of this particular conversation, please send me names and email addresses so I can reach out to people.
>
> A potential deliverable out of a dedicated MathML sub group could be a position paper which enumerates the use cases and vets some of the various solutions that people have been considering across the board.
>
> What do you guys think? I’m happy to push some organization behind our combined conversations in order to keep momentum going.
>

I do believe that such a study would be useful; what I do not know is whether the MathML group has, or has planned to, do something similar. Luckily, Bert Bos is in our midst, and he is also staff contact for the MathML WG, so he can tell us.

> Why do I want to do this, and for those of you who don’t know me, where the heck did I come from? I was WAY more hands on involved with MathML back in 2006-2007 when I worked for Cengage and was asked to look into some of the thorny MathML problems and existing solutions of the time. There’s been some progress since 2006-2007, and I’ve changed roles a couple of times, but the MathML problems are still here, and include new and different use cases than were previously identified. I don’t work for a publisher any more, but I am currently back in a role where I can start advocating for MathML solutions again, so here I am.
>

What we should also keep in mind is the charter of this specific IG. Although we have the liberty to decide the publication of such study, our primary goal is to identify possibly missing features of the OWP. Let us suppose we were in an ideal world and all rendering engines had a full MathML implementation (a guy can dream:-): would that study still be of importance? I ask that because we have to give a clear scope for what we are planning here...

A related question: I know the Readium project aims at providing a software _on top_ of current rendering engines. I also know that there was a MathML module bound the WebKit for a while but it has been taken out:-(. Isn't it, wouldn't it be possible for the Readium project to implement a MathML module once and for all and thereby removing this thorn from the publishers' side?

Ivan

> Please let me know any questions, comments, further ideas for discussions, or other things that you think can come out of such a sub group. I won’t be answering further emails on the topic until after 2 pm EDT today. I need to go finish my slides for today’s webinar now, or Peter and Sanders are going to have a lot more on their plates than they signed up for! ;)
>
> Thanks for your attention and for continuing the conversation.
>
> Jean Kaplansky
> Solutions Architect
> Aptara, Inc.
> Email: jean.kaplansky@aptaracorp.com
> Skype: JeanKaplansky
> Mobile: 518 487 9670
>
> <ACD331E8-87DF-4649-A68D-1AC6C320CC10[45].png>
>
> From: Richard Schwerdtfeger <schwer@us.ibm.com>
> Date: Wednesday, November 6, 2013 at 9:59 AM
> To: Gerardo Capiel <gerardoc@benetech.org>
> Cc: Ivan Herman <ivan@w3.org>, Jean Kaplansky <jean.kaplansky@aptaracorp.com>, Peter Krautzberger <peter.krautzberger@mathjax.org>, W3C Digital Publishing IG <public-digipub-ig@w3.org>, W3C Public Digital Publishing IG Mailing List <public-digipub-ig-comment@w3.org>
> Subject: Re: The AAP EPUB 3 Implementation Project
>
> yes.
>
>
> Rich Schwerdtfeger
>
> <graycol.gif>Gerardo Capiel ---11/06/2013 08:30:29 AM---To be a bit more specific, MathJax can output SVG when it runs in the user agent (e.g., a web browse
>
> From:	Gerardo Capiel <gerardoc@benetech.org>
> To:	Richard Schwerdtfeger/Austin/IBM@IBMUS
> Cc:	Jean Kaplansky <Jean.Kaplansky@aptaracorp.com>, Ivan Herman <ivan@w3.org>, Peter Krautzberger <peter.krautzberger@mathjax.org>, "W3C Digital Publishing IG" <public-digipub-ig@w3.org>, W3C Public Digital Publishing IG Mailing List <public-digipub-ig-comment@w3.org>
> Date:	11/06/2013 08:30 AM
> Subject:	Re: The AAP EPUB 3 Implementation Project
>
>
>
> To be a bit more specific, MathJax can output SVG when it runs in the user agent (e.g., a web browser or eReader). However, as Rich mentions, this process of converting MathML or other math input formats to browser renderable formats in the user agent does require runtime processing and would not be as performant as converting the MathML to accessible/described SVG prior to EPUB packaging and/or transmission to the user agent.
>
> Gerardo Capiel
> VP of Engineering
> Benetech
>
> On Nov 6, 2013, at 7:15 AM, "Richard Schwerdtfeger" <schwer@us.ibm.com> wrote:
>
> Hi Jean,
>
> No. We can use ARIA for either HTML or SVG. SVG is just a better fit for mobile devices and is a lot simpler. I have been discussing the SVG approach with Gerardo and he has covered that topic in his post. MathJAX currently renders MathML using JavaScript and HTML but it slower. SVG was designed for drawing.
>
> ARIA semantics, currently, are largely designed for widgets and landmarks. We could use labelling to provided accessibility information to the content either through <title> or <description>. We could also create a taxonomy around the Math symbols that are drawn. I fact, we have discussed creating other modules for ARIA (instead of one big monolithic spec.) to deal with other taxonomies - charts, math, etc.. The tabindex support will be big for Math rendered in SVG.
>
> I will try to attend the webinar today. Thank you for the pointer to the use cases.
>
> Rich
>
>
>
>
> Rich Schwerdtfeger
>
> <graycol.gif>Jean Kaplansky ---11/05/2013 06:13:27 PM---Hi Richard – Welcome to the group! I just saw your comments in my inbox, and as I’m getting ready to
>
> From: Jean Kaplansky <Jean.Kaplansky@aptaracorp.com>
> To: Richard Schwerdtfeger/Austin/IBM@IBMUS, Ivan Herman <ivan@w3.org>, Peter Krautzberger <peter.krautzberger@mathjax.org>
> Cc: W3C Digital Publishing IG <public-digipub-ig@w3.org>, W3C Public Digital Publishing IG Mailing List <public-digipub-ig-comment@w3.org>
> Date: 11/05/2013 06:13 PM
> Subject: Re: The AAP EPUB 3 Implementation Project
>
>
>
>
> Hi Richard – Welcome to the group!
>
> I just saw your comments in my inbox, and as I’m getting ready to do a mathml webinar tomorrow, you immediately got my attention…
>
> You say that “A big issue will be getting open source MathML readers (voice) and renders – preferably based on SVG as it is a good fit for mobile.”
>
> Do you mean that it would be better to have MathML that is rendered as SVG than rendered natively in a browser because of the ARIA and tab index support?
>
> The DPUB has started to collect some MathML use cases here: http://www.w3.org/dpub/IG/wiki/MathML_UC. We would definitely like to identify and record additional use cases around MathML and Accessibility. Please go ahead and send email to the group regarding any MathML use cases you wish to add. I’ve been the “de facto” keeper of the MathML use cases for the group to date. (yep. I volunteered. And sort of dragged Peter along with me, since he’s the guy who provided the initial write-ups of the 3 use cases we’ve documented to date…)
>
> Here’s the link to register for the webinar we’re doing tomorrow if anyone’s interested:
>
> https://event.on24.com/eventRegistration/EventLobbyServlet?target=registration.jsp&eventid=681450&sessionid=1&key=BF15E182E16C4F1C3A04C885DF1DC212&sourcepage=register
>
> This was originally going to be a 2 person effort between Peter Krautzberger of MathJax and myself, but we recently expanded our speaker list by one to include Sanders Kleinfeld from O’Reilly who will be presenting a case study about how O’Reilly built a toolchain API to deal with the need to put math on different retail reading devices in addition to mobile apps, the web, and in printed books.
>
> Jean Kaplansky
> Solutions Architect
> Aptara, Inc.
> Email: jean.kaplansky@aptaracorp.com
> Skype: JeanKaplansky
> Mobile: 518 487 9670
>
> <2A579430.gif>
>
> From: Richard Schwerdtfeger <schwer@us.ibm.com>
> Date: Tuesday, November 5, 2013 at 6:57 PM
> To: Ivan Herman <ivan@w3.org>
> Cc: W3C Digital Publishing IG <public-digipub-ig@w3.org>, W3C Public Digital Publishing IG Mailing List <public-digipub-ig-comment@w3.org>
> Subject: Re: The AAP EPUB 3 Implementation Project
> Resent-From: <public-digipub-ig-comment@w3.org>
> Resent-Date: Tuesday, November 5, 2013 at 6:57 PM
> Ivan,
>
> This is a great. Going forward we will be able to expand on ePub's accessibility capabilities. For example, SVG 2 now has ARIA and tabindex in the spec. In the future we will be able to roll this support into books. A big issue will be getting open source MathML readers (voice) and renders - preferably based on SVG as it is a good fit for mobile. Access to digital math by the blind and those with attention deficit would benefit from having this capability. In fact we recently did a study for the Department of Education. The ability for students to have Math navigable through highlighting and speech improved math comprehension by 10-15 percent across the board.
>
>
> Rich
>
>
> Rich Schwerdtfeger
>
> <graycol.gif>Ivan Herman ---11/05/2013 05:34:09 AM---I would expect many of you (most of you?) may be familiar with this, but gor those of you who are no
>
> From: Ivan Herman <ivan@w3.org>
> To: W3C Digital Publishing IG <public-digipub-ig@w3.org>, W3C Public Digital Publishing IG Mailing List <public-digipub-ig-comment@w3.org>
> Date: 11/05/2013 05:34 AM
> Subject: The AAP EPUB 3 Implementation Project
>
>
>
>
> I would expect many of you (most of you?) may be familiar with this, but gor those of you who are not:
>
> https://nfb.org/images/nfb/documents/html/aapepub3implementation.xhtml
>
> Lots of good guidelines point on accessibility issues, for example
>
> Ivan
>
> ----
> Ivan Herman
> +31 641044153
>
> (Written on my mobile. Excuses for brevity and frequent misspellings...)
>
>
>
>
> [attachment "graycol.gif" deleted by Richard Schwerdtfeger/Austin/IBM] [attachment "graycol.gif" deleted by Richard Schwerdtfeger/Austin/IBM] [attachment "2A579430.gif" deleted by Richard Schwerdtfeger/Austin/IBM]
>
> <graycol.gif>

----
Ivan Herman, W3C