project page


About gNumerator
General Info
Plans and Progress
About MathML
SourceForge Project Page for gNumerator
Updates / News / Blog
Blog, Updated January 26, 2005
Downloads, release 0.34,
Feburary 13, 2005
Release Notes for version 0.34
MathML Control, MathML DOM, Test App, and Documentation
Source Code
gNumerator Mailing List
Need an experienced computer scientist?
My Resume / CV
Anonymous Feedback
Anonymous Feature Requests
Anonymous Bug Reports
Extending the DOM
Fast parsing of XML entities
Calling C methods from C#
The visitor pattern
MathML Rendering Control
First Rendering of MathML
A slightly more complex example
Same example, with almost correct use of stretchy characters
Stretchy operators are finally working correctly
SubScripts and SuperScripts are now working, also a comparison with Mozilla
Radicals are now working, and a comparison with Microsoft Word
Radical Indices, and sub and super scirpt elements
Under and Over Elements, and Horisontal Stretchy Operators
Complex Tables
Browse Source Repository
MathML Document Object Model
MathML Rendering Control

Square Root elements are now working! This includes <msqrt>, and <mroot>elements. I used the CMEX10 (computer modern math extension) font for the square root glyphs. The nice thing is that this font provides many glyphs for this character, and each glyph is angled slightly differently, so they look good with various sized radicands. Also, this font provides a true stretchy glyph for radicals that must stretch to very large sizes. Notice how using differnt glyphs produces much better results than Microsoft Word (screenshot at bottom) which only used the Windows font mapper to stretch a single square root glyph to different sizes.

A note about fonts: Most glyphs are rendered from the standard "times new roman" and "symbol" fonts, as they come installed in all Windows systems. I currently use the TeX CMEX10 font for radical glyphs, and probably will use it for more glyphs in the future, as this font is the only freely availible font that can produce these glyphs properly. The Mathematica fonts also produce these glyphs, but they are not freely available, or at least I am not legally allowed to redistibute them.
All glyphs mapping (in the rendering widget) is handled via a set of config files. Each config file (standard xml) is for a single font, so currently, I have "font-configuration-times.xml", "font-configuration-symbol.xml", and "font-configuration-cmex.xml". The rendering widget enumerates the contents of the config directoy at startup time, and all font config files are loaded. I probably will create a config file for the Mathematica fonts, and if a user wants to, they can download this font, and the rendering widget will automaticlly load it. Note, the Mathematica fonts do produce better looking glyhps than the TeX computer modern font.

source mathml for this screenshot

(above) A screenshot of the MathML rendering control showing square roots, and stretchy integral operators. Notice the different sized glyphs for the square root symbols, as they encompass different sized radicands. Also, notice how the integral operators stretch properly to the height of the row that they are in. I use the standard symbol.ttf font for the integrals and parentheses, as it provides good multi-part glyphs for these, and I use the CMEX10.ttf font for the radicals. The other big difference in this rendering, and the Word equation editor (below), and the Mozilla rendering of MathML is that the gNumerator MathML rendering widget produces much better, tighter fitting renderings. This is because it uses the Padovani area fitting algorithm to do the fitting and layout.

(above) A screenshot from Microsoft Word's equation editor. This is not the same MathML as the previous example, as Word's equation editor can not do mathml, it has some proprietary data format. Notice how poorly the radical glyphs 'stretch'. This is because they are using a single glyph for the radical, and using the Windows font mapper to 'stretch' it. I originally played around with this technique, but abandoned it very soon, as I produced extremly poor looking glyphs.