MathML amidst Open Web Standards: Mozilla's Building Blocks for Today and Tomorrow
Roger Sidje
The University of Queensland
Abstract
The story of MathML goes back to the early days of the web itself. After years of vigorous debate and search for consensus in ways to publish mathematics through the web, there was a sense of relief when the first MathML recommendation was announced in 1998. But it soon appeared that traditional browser makers were not in any particular hurry to invest their resources into integrating its support. That was the beginning of another wait. Although third parties started offering stop-gap MathML tools, the spotlight really never went away from mainstream browsers, and the scientific community nurtured its patience. The alternative of proprietary solutions didn't appeal to the majority of users and as a result, MathML content didn't spread very much on the web.
Over recent years however, there have been concomitant efforts by third parties which today are resulting in readily available means to represent MathML in the major browsers. For instance, recent versions of Microsoft Internet Explorer can render MathML in the flow of the main text using the MathPlayer pluggin developed by Design Science (maker of MathType). While the usage of MathPlayer involves non standardized markup elements, the details can be shielded and users can cater for all browsers by using a XSLT StyleSheet specifically made by the MathML Working Group for that purpose (dubbed Universal Math StyleSheet). The MathPlayer is an example of a new generation of tightly integrated pluggins based on the so-called "behavior" or "binding" technique. On the other hand, MathML support is provided natively in Amaya and Mozilla (the source code for Netscape 6.x).
The availability of MathML in mainstream browsers is a significant turning point, marking the beginning of a new era for representing mathematics and science on the web. It was the firm conviction of the MathML Working Group that a standardized language would significantly enhance interactive scientific communications. That conviction is still a strong motivating force and is fueling the various efforts at implementing and integrating MathML as a glue language cross-linking browsers, computational/numerical kernels and/or reasonning/symbolic engines such as Mathematica or Maple. However, several challenges still lie ahead as MathML is not intended to be an end in itself. There are several open issues to keep the MathML community busy for the foreseeable future. These include:
- Rendering/Representation
While there has been significant progress with MathML renderers, other aspects such as bi-directional rendering (i.e., for non-Latin languages such as Arabic or Hewbrew) still need to be integrated in the MathML specification and then in the renderers. MathML exchange (e.g., e-mailing) raises other questions. - Authoring/Production
The very nature of MathML requires automatic generators, convertors, and MathML-aware editors. Much progress is being made but improvements are needed. Other tools and utilities are needed for the production and migration of legacy content towards MathML-aware environments. - MathML Mining
Cross-referencing, indexing formula/proofs, and searching tools. Exploring and discovering knowledge in MathML archives still need thorough investigation. - Interoperability
W3C has provided an array of recommendations (MathML, XHTML, XML, DOM, CSS, XSLT, SVG, SOAP, etc) that are intended to interact with each other. These are culminating into what is termed web services. Elsewhere, interfacing MathML with other applications such as Mathematica or Maple is well underway although using proprietary solutions.