## Mozilla MathML Project: status report

By fredw on Tuesday, January 10 2012, 19:24 - Permalink

So here is the usal review of activities in the Mozilla MathML project during the last six months:

- I finished the work started last summer by Jonathan Hage regarding the implementation of the align attribute on <munder/>, <mover/> and <munderover/> elements (Gecko 12.0).
- As said in my previous report MathML linking is available again since Firefox 7. However, the context menu still has some issues with MathML3 href. Marco Bonardo and Bill Gianopoulos has continued the work but it is not finished yet. See bug 534968 if you want to help. Boris Zbarsky also fixed a bug with color of visited/unvisited MathML links (Gecko 10.0).
- Implementation of
`<maction/>`

has been cleaned up (Gecko 9.0 and 11.0), in particular the non standard actiontype "restyle" is removed. Masayuki Nakano improved how click events are handled on that element. `<mlabeledtr/>`

is still not implemented but this element becomes more important now that MathJax 2.0 is going to rely on it for equation labeling. Firefox did not display at all such a row, but it has been suggested since a long time to be less strict and to just hide the label instead. That's now done (Gecko 9.0). The bug for a complete implementation is bug 689641.- I've mentioned this feature several times in the past, but I finally finished the
implementation of right-to-left writing of mathematical formulas (Gecko
12.0). I would like to thank Azzedine Lazrek, Mustapha Eddahibi and Khaled
Hosny for their helpful remarks on Arabic notations. Here is an example of
formula that should render right to left in Firefox 12:
$$\sqrt{\sum _{\u066e=1}^{\u0635}{\u0633}^{\u066e}}$$

Any volunteer from the Mozilla Arabic community to try it more deeply and/or write demo pages?

Among future improvements, I can mention code clean up, such as removal of unused MathML atoms by Florian Scholz or a refactoring of attribute parsing that improves conformance against the MathML3 recommendation. I have also started implementation of <mglyph/>. Thanks to discussion with Davide Cervone (the main author of MathJax) I was able to convert MathJax tables for stretchy and large operators into our own format. That means that we will be able to use MathJax fonts which look closer to TeX and thus could improve the rendering of formulas in Firefox. To give an idea, here is a screenshot of the MathML torture test:

Left: TeX rendering

Right: Firefox rendering, with experimental MathJax fonts support