Re: Recent MathML Improvements in WebKit

On 21/07/2016 07:00, Frédéric WANG wrote:
> Hi everybody,
>
> People following native MathML developments may know that some work has
> been done in WebKit by Igalia developers in the past few years. Igalia
> [1] is a free software consultancy with experts working on all open
> source web engines (maybe some of you have heard about developments in
> e.g. CSS Flexbox and Grid layout, ECMAScript, WebRTC, Streams API or
> ARIA). In the past, I got the opportunity to collaborate with some
> Igalia employees. In March I joined Igalia's web platform team.
>
> During my first months at Igalia I have been in charge of upstreaming
> many of the MathML patches and to continue the improvements.
> Essentially, all the design issues of the flexbox-based implementation
> raised by reviewers in the past are now addressed. And although there is
> still some work to do and known bugs, we believe we have made some
> significant improvements in the MathML support and rendering quality.
> People interested can find details on by blog [2] [3].
>
> We are very interested to hear any feedback from MathML enthusiasts [4]
> that could help future developments, for example those that will be made
> during the next Web Engines Hackfest in September [5]. OS X users will
> be able to test the new features in the next Safari Technology Preview
> [6]. Linux users will have to build WebKitGTK+ themselves [7] or
> otherwise wait the next release of GNOME in september and the
> integration in their respective distributions. As usual, you must have
> math fonts installed on your system to get proper rendering [8]. In
> addition, you can add a few CSS lines to your user agent stylesheet in
> order to test it on Wikipedia [9].
>
> We also have feedback to provide to the authors of the MathML
> recommendation from the point of view of web engine developers
> (including many similarities with what was reported for Gecko in the
> past). It seems that the Math WG is now inactive but hopefully comments
> can be taken into consideration for future documents. If members of the
> former Math WG are interested to improve the compatibility and
> consistency with the web browser code base, then we will provide this
> information in follow-up messages.
>
> The most important point is that the MathML in HTML5 implementation note
> [10] and its accompanying test suite [11] from the MathML Association
> were very important to write a good MathML implementation in WebKit and
> to have MathML features verified in the integration testing framework.
> Volunteer developer James Kitchener has also relied on it to continue
> improvements in Gecko and to add many tests to Mozilla's own framework.
> Several web engine people have been enthusiast about these
> implementation note and test suite and suggested to move them into an
> official Math WG's document and to the W3C's web-platform-tests [12] so
> that all the web engines developers and specification authors can work
> on it.
>
> We believe that with the experience in Gecko and WebKit as well as our
> experimentation in Blink, we can now be confident enough on this
> document and tests to move it under the W3C's umbrella and continue
> improvements there. The implementation proposed by the note relies on
> TeX rules and on the OpenType MATH table and some Microsoft employees
> seem inclined to consider that approach for Edge too [13]. So again we
> would really like to know if any members of the former Math WG are
> willing to collaborate with web engine developers on this note, test
> suite and hence help native MathML implementations?
>
> Frédéric, for the Igalia's Web plaform team.
>
> [1] https://www.igalia.com/
> [2] http://frederic-wang.fr/mathml-refactoring-in-webkit.html
> [3] http://frederic-wang.fr/mathml-improvements-in-webkit.html
> [4] https://bugs.webkit.org/enter_bug.cgi?product=WebKit&component=MathML
> [5] http://www.webengineshackfest.org/
> [6] https://developer.apple.com/safari/technology-preview/
> [7] https://trac.webkit.org/wiki/WebKitGTK
> [8] http://trac.webkit.org/wiki/MathML/Fonts
> [9]
> https://www.mediawiki.org/wiki/Extension:Math/advancedSettings#CSS_for_the_MathML_with_SVG_fallback_mode
> [10] http://www.mathml-association.org/MathMLinHTML5/
> [11] http://tests.mathml-association.org/
> [12] https://github.com/w3c/web-platform-tests
> [13] http://lists.w3.org/Archives/Public/www-math/2016Apr/0014.html,
> https://twitter.com/SampsonMSFT/status/727199790736384001
>
>

I've been following recent webkit bug activity and this is really
excellent news, thanks for the work that's going into webkit.

While it's true that the math WG is currently "on hold" as there were no
planned date to do a MathML4 to give implementations time to catch up
with MathML3, this mailing list is still active and specific suggestions
can be posted here. Once it's clearer exactly what documents
would make sense to move to W3C it's certainly not impossible that
some activity could be proposed to be re-chartered to enable that.

David

________________________________


The Numerical Algorithms Group Ltd is a company registered in England and Wales with company number 1249803. The registered office is:

Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.



This e-mail has been scanned for all viruses by Microsoft Office 365.

________________________________

Received on Thursday, 21 July 2016 06:53:10 UTC