W3C home > Mailing lists > Public > www-math@w3.org > February 2013

RE: Help get math turned back on in Chrome

From: Paul Topping <pault@dessci.com>
Date: Thu, 7 Feb 2013 22:46:01 +0000
To: Dave Barton <dbarton@mathscribe.com>, "www-math@w3.org" <www-math@w3.org>
Message-ID: <B6C5B1ABA88AF446821B281774E6DB7107832D@FERMAT.corp.dessci>
Hi Dave,

I think most of us agree that Google has the right to say the code is not production ready. It is the lack of an explanation that hints at what needs to be fixed that is objectionable. We can perhaps guess by looking at some list of bugs but I am sure every part of WebKit/Chrome has unresolved bugs that don't stop code from being production-ready.

While you have done great work in adding MathML to WebKit, I don't think it is sufficient to break the logjam you refer to. Most publishers have to be absolutely sure that virtually all of their readers can read what they produce. To break the logjam, you would have to add robust MathML support to ALL web browsers and ebook readers. MathJax has a better chance at breaking the logjam as it can promise MathML support to all web browsers. This allows publishers to start using MathML thereby making the chicken from scratch (or is it the egg?).  Once that happens, people may note that having MathJax do the MathML rendering is not quite as good as native code in the browser and this will drive bringing it inside the browser. (This is not meant as any kind of slam against MathJax. If people find that it does well enough, perhaps there will be no need to bring it inside the browser.)

While I understand your argument for ebook readers to move from WebKit to Firefox, I doubt that any ebook reader maker would justify such a move solely because of better math support. Because ebooks are paginated and ebook reading systems more constrained than the web page environment, the problems of MathJax vs native MathML support will be less apparent in ebooks.

I think we can poke Google without being seen as troublemakers. That said, I can say from experience that one software company or community telling the bigger software company that they should do something will fall on deaf ears. If Google sees that the demands for MathML support all come from the MathML community, they will have good reason to disregard it as the usual ranting of some group trying to push their favorite technology. Big companies get so much of this that they routinely ignore it.

For the campaign to be successful, it has to come from their customers. If education and/or accessibility communities demanded MathML support loud enough, I believe it would force them to respond. Another way in is to find specific people high up in the organization to champion our cause. They are much more likely to listen to someone inside.

One thing in favor of our campaign is the relatively low cost of solving it. That's what is so frustrating to us. Google, Microsoft, and Apple each have so much cash that putting a few developers on solving this problem would not even show up on their budgets.

Paul

From: Dave Barton [mailto:dbarton@mathscribe.com]
Sent: Thursday, February 07, 2013 1:58 PM
To: www-math@w3.org
Subject: Re: Help get math turned back on in Chrome

On Wed, Feb 6, 2013 at 3:11 PM, Carl Malartre <malartre@scolab.com<mailto:malartre@scolab.com>> wrote:

Please be respectful of the Chrome folks. What's the proof that they are dishonest and not caring?

Development is hard and manpower is finite and they have to select what they support. They want a good product and I think they have the right and authority to say "the code is not yet production ready".

They said "We hope to turn it on in some future release.". You don't believe them?

On Feb 7, 2013, at 6:14 AM, Daniel Marques wrote:
It is clear that users and the industry in general need MathML in the browser. Let's be supportive with the chromium team and if they say that it will be available in a future release, well, just say that we believe them.

In the meanwhile we should help in improving the MathML implementation in Chrome and do all possible to convince them that it is worth to put it back as soon as possible.

This is a very important discussion, possibly critical to MathML. I really think we should hash this out here so everyone understands what's at stake.

I worked as a volunteer for a year with the webkit (Chrome) developers, especially the ones at Google. They are not stupid or evil; some I consider good friends. However, Google is very powerful, and frankly a bit insular at times.

The historic problem with MathML, as Ian Hickson (chief HTML5 editor) and others have pointed out, is the chicken-and-the-egg problem: people don't use it in web pages because it doesn't have good browser support, so browser vendors don't bother supporting it because it's not in a lot of web pages. Refining that, if Chrome waits for Internet Explorer, and Internet Explorer waits for Chrome, then neither will get it. In this state, it's hard for web page authors to rely on MathML support. To help break this logjam is the reason I volunteered.

Google has lots of Chrome developers working on features for web designers and graphic artists. They, and the rest of the webkit companies, don't have even a fraction of a person assigned to MathML. While Google may hope to turn MathML back on in the future, at this time they are not working to make that happen.

One reason for hope is digital textbooks (e-pub). These will soon need a good MathML implementation. My understanding is that the reference implementation for e-pub uses Chrome and a javascript library (MathJax). Such a library can emulate MathML, but for may reasons (speed, integration with CSS and javascript and localization/SVG/etc.) it's much better to implement MathML natively in the browser, and just use the javascript library to work around limitations. For this reason, should e-pub work shift from using Chrome/webkit to using Firefox?

Look, I'm not making this suggestion as some kind of a threat. It's an honest question, and there are experts on this list that can answer it much better than I. (I might also mention that there are experts on this list, including developers of other MathML implementations, that would freely provide advice or in some cases even code to webkit/Chrome.) The truth is that none of the webkit/Chrome developers really has any knowledge or interest in MathML, so frankly they just see it as putting a burden on the whole project. But the MathML code is small, currently about 2,000 lines not counting copyright notices, and a fraction of a developer could easily maintain and slowly improve it.

This is the w3c mathml mailing list. I think it is appropriate that we discuss here what it means for MathML if Chrome does not support it, and frankly what that means for Chrome. We need to pick a way forward, and when we achieve clarity I suggest we all go back to Neil's e-mail at <http://lists.w3.org/Archives/Public/www-math/2013Feb/0005.html>, link to it, and publicize it. In the next few years, literally hundreds of millions of students, scientists and engineers will want to view mathematics natively in web browsers. But if lots of people don't star the chromium issue as Neil suggests, Google will conclude from the lack of MathML in web pages that it really isn't a priority after all, and there simply will be no resources put on it.

Hope this helps, Dave B.
Received on Thursday, 7 February 2013 22:46:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 7 February 2013 22:46:33 GMT