I think the main points were: 1) Not everybody use TeX as an input method. 2) Not everybody write the source of Web pages by hand. 3) People using TeX want its full power (defining macros, loading packages etc).
So except in simple and limited cases, a small subset of TeX is not what people want. MathML already contains the core features for mathematical rendering and has been widely used for a long time. Concrete tools have been developed and shown that MathML can be used with other Web languages (HTML, SVG and CSS), with DOM/Javascript, to write mathematical search engines, can be generated from AsciiMath, can be used for accessibility (*), in WYSIWYG editor, in ebooks, for copy and paste (e.g. Microsoft Word, Mathematica/Mapple), in publishers's internal workflow etc For the LaTeX community, tools like LaTeXML (currently able to convert 94% of the arXiV papers) have shown that MathML can be generated from LaTeX and that the <annotation> element can even be used to expose and share the original LaTeX markup (e.g. for copy and paste). So your proposal of a small subset of TeX is mostly (depending on what you put exactly in the subset and how you map it to the DOM tree) a syntactical change, just like the RelaxNG XML form vs the RelaxNG compact form. As I said, I understand this can be helpful when one wants to quickly write a HTML page with math content and Javascript libraries like MathJax or AsciiMath have already addressed that need. With non-XML/SGML syntax, you're just making more difficult for authors/implementers to understand what will be the DOM tree you get at the end. So you're making life easier for people writing their Web pages by hand with no CSS or Javascript involved, but much more complicated in all the other cases. You claimed that MathML was not used and that replacing it by the more universal TeX syntax would make developing tools easier (like accessibility or copy and paste) while things that do not work well with TeX (like CSS) were not absolutely necessary. That just seemed misinfor mation about the current status of MathML and current use cases. You asked advice about the usefulness of MathML on the MathJax list and we tried to explain that it is indeed very important. Of course TeX is important too, but is not always adapted to a Web context (remember it was initially designed by Knuth to write books). As I said, I would prefer improving MathML, especially compatibility with CSS, than starting again from zero with a new math language for the Web and reinventing the wheel. (*) BTW, ChromeVox has recently been released with MathML support. So Google can now read the math but not display it: http://www.youtube.com/watch?v=YyWu9HB9QtU _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
