- From: <juanrgonzaleza@canonicalscience.com>
- Date: Mon, 22 May 2006 06:17:56 -0700 (PDT)
- To: <www-math@w3.org>
Once solved the downloading problems with Mozilla fonts for MathML, I believe that this would be a good moment for reviewing the rendering-math-via-special-fonts philosophy. The Web is not done of paper and whereas a paper is distributed in a specific font (chosen when printing the document), you never can be sure if fonts you are using are available on the client side. Moreover, one of basic principles of accessibility is that web pages should be platform independent, meaning that they should be accessible regardless of the user’s platform and settings. Those general guidelines are violated via forcing the download and use of special fonts for mathematics. It would be really strange the rendering of HTML tables via forcing to users download and installation of special fonts for tables! One of traditional obstacles for the widespread acceptance of TeX in the web has been that TeX heavily relies on special fonts were not designed for the Web. Computer Modern fonts can be not readable on the screen for certain combinations of sizes and resolutions. I find just curious the, at the one hand, claim that TeX is not designed for the Web whereas, at the other hand, one relies on TeX fonts for "correct" rendering of MathML formulae. To rely on specific fonts rendering is, in my opinion, contrary to original MathML goal of serving mathematics in an independent device manner. Therefore, I find a bit disturbing that none folk has noted this here. I just waited something as "download and install specific fonts for Mozilla if you want but have care with this and this..." >From L. Padovani <blockquote> Mozilla relies on specific fonts being installed on the system in order to work correctly. In particular, it assumes that Computer Modern fonts are available. This means that the formatting engine is polluted with dependencies on Computer Modern fonts. On one side it is true that these fonts are some of a very few fonts for mathematics that are currently available, hence committing to them is not really ruling out alternatives (nevertheless we have to remember Mathematica and Euclid fonts for mathematics). Needless to say, basically no fonts other than Computer Modern adopts the same conventions, which basically means that MathML support in Mozilla is extensively tied to these fonts. </blockquote> and adds <blockquote> According to the STIX home page, a brand new family of fonts for mathematics will be available in the next future. This will imply a significant rewrite of the MathML formatting engine of Mozilla (note that it is not possible to just replace a component: the whole application must be recompiled). </blockquote> Finally, I would recommend to users a reading of fonts licenses available on Mozilla site before installing them and also perform some rendering tests (some MathML documents take from 10 to 20 minutes to render in Mozilla when others approaches take of order of 2 seconds). Of course, there exist other approaches to render mathematics on the web are not based in specific fonts and fit with general CSS guidelines (e.g. the famous users’ stylesheet settings). Juan R. Center for CANONICAL |SCIENCE)
Received on Monday, 22 May 2006 13:18:13 UTC