MathML Subcommittee teleconference minutes:

On Wednesday 7/22, the MathML subcommittee held a conference call to explore what type of action (if any) the WAI-UA group should take concerning access to MathML. Below are the minutes from that meeting. Sorry for the late posting.

WAI-UA-MathML subcommittee teleconference minutes 7/22/98

Tom Wlodkowski
Madeleine Rothberg
Jim Allan
John Gardner

Kinds of user agents: 
1. All agree that it will be difficult for a screen reader to interpret MathML in a standard visual rendering. Two possible solutions are
1) a specialized audio browser that reads the MathML code and renders it directly to audio. Should be written by someone with understanding of the needs of people with visual impairments.
2) render the MathML visually in a linear way, with special symbols for the different operations. This is the approach taken by TRIANGLE. Once the math is linearized any screen reader could read it provided the screen reader  was equipped to pronounce the extra symbols.

2. It will likely be a long time before MathML rendering is a part of the mainstream browsers, which means rendering will be achieved through plug-ins or applets Either of the two solutions mentioned above could be implemented this way.

Potential UA-MathMl guidelines:

1. Once browsers are capable of rendering MathML, they should provide access to all controls needed to configure it's use.

2. Use of plug-ins may require using MIME types. Users need to be able to select correct types so that MathML is rendered in their preferred plug-in (i.e. Math/linear or Math/audio for example).

Semantic and presentation markup:
1. Madeleine Rothberg suggested that access for blind and visually impaired people would be much better from semantic markup.

2. Perhaps, when possible, the semantic information could be available if you asked for it, like pop-up tool tips.


1. authoring tools and page authors should be encouraged to use semantic markup as much as possible since it will be useful to a broad range of users, including people with disabilities.

2. There is a need to figure out how a screen reader might interact with an audio browsing plug-in. As you read down a page, how is the handoff made when you get to the math? How is control passed between the plug-in and the screen reader?

3. The Document Object Model (DOM) may be helpful for MathML access, once it is finalized. We'll need a DOM for the math objects that properly defines the objects so that you can browse them and understand everything.

4. A discussion of how braille access to MathML can be achieved. John Gardner mentioned that the general concensus of the braille working group is that the braille output need not match the screen exactly. A translator that had programatic access to the MathML code or DOM could handle the translation.

Next steps:
1. In order for the UA to develop guidelines for dealing with MathML, a renderer must first be built and the DOM needs to be finalized.

2.John Gardner testing out Amaya which has some MathML support. (Some Discussion concerning Amaya has already taken place on this list since this meeting).

3. Conversation to be continued via e-mail.

Received on Thursday, 30 July 1998 17:44:09 UTC