Minutes – Z39.86 Advisory Committee Conference Call

Attendees

Advisory Committee:

• George Kerscher, Advisory Committee Chair – DAISY Consortium and Recording for the Blind & Dyslexic
• James Pritchett – Recording for the Blind & Dyslexic
• Lloyd Rasmussen – National Library Service for the Blind and Physically Handicapped
• Marisa DeMeglio – DAISY for All Project
• Neil Bernstein – National Library Service for the Blind and Physically Handicapped
• Ole Holst Andersen – The Danish National Library for the Blind
• Thomas Kjellberg Christensen – Danish Library for the Blind
• Linus Ericson, Swedish Library of Talking Books and Braille
• Laurie Sherve, Scribe – DAISY Consortium

MathML Working Group

• Jonathan Williford, GH
• Dennis Leas, GH
• Emilia Persoon

Participating Experts:

• Chris Wilder-Smith – CAST
• Neil Soiffer – Design Science
• Dominic Labbé – HumanWare

Apologies:

• Dave Pawson – Royal National Institute of the Blind
• Mark Hakkinen, Daisy For All Project
• Markus Gylling – DAISY Consortium and Swedish Library of Talking Books and Braille

Agenda

The agenda was distributed to the Advisory Committee's list on 2006 January 03. There were no amendments in advance of the call.

There were no additions to the proposed agenda.

Confirmation of Date and Time for Next Call

GK proposed that the next call be held on 2006 January 25.

Overview and approach - MathML

The group decided to consider four cases:

1. a native MathML player capable of both visual rendering;
2. a native MathML player capable of visual rendering with audio/text fallbacks
3. a native MathML player capable of audio rendering with an image fallback
4. a non-MathML aware player that uses audio and image fallbacks

A native MathML player may also be capable of braille math code generation. Unlike visual math notation which has a widely used notation, there are many different braille math codes used around the world. These codes differ from a straight translation of the spoken math.

XML Discussion

The proposal adds two new elements: "extensiongroup" and "extension". Extension has a "type" attribute to specify an extension, such as the "URI for MathML extension".

There was generally positive feedback about the additions; it is not clear if they will work for video since there are questions as to whether or not the SMIL file overrides info in the XML file. That question was tabled for another day.

James feels that our current Z39.86 spec extension mechanism will allow this to be brought into the DTD via internal subsets. If we want to build this into the DTBook vocabulary, it would be a spec revision that would require re balloting. The consensus was that a spec revision was not desirable.

"extension" and "extensiongroup" would be an independent (of the math extension) portion of the recommendation for others (including math) to use. It would make use of a DAISY namespace URI.

Braille issue

There was a long discussion concerning the braille in DTBook rather than a not <showinbraille> and text. It is possible to have simultaneous alternate displays of just braille, just text, or both braille and text. Marisa will bring the issue of two simultaneous text representations to the SMIL working group. The Zed committee will address this issue when we address the submitted issue about <showin> attributes.

SMIL discussion

The proposal included the use of "switch", which is not part of the 2005 spec and the use of the "ref" attribute which is also not part of the 2005 spec.

James suggested not using "switch" but the attribute <customTest> which was envisioned for use with "skippable structures". This (along with the removal of ref [see below]) would make the SMIL file compatible with the 2005 spec. James walked through how this might be done by defaulting the tests for the three MathML cases to "false" and the 'dumb' case to true. All players that understand <customTest> should handle this properly.

Thomas suggested a mechanism for grouping a set of custom tests so that the cases act like radio buttons (only one can be true). People agreed that this would be good in a Z39.86-next, but it is not essential now. A 'dumb' player would not recognize the three MathML tests, and MathML-aware players can enforce the "only one" behavior (they would probably pick the case they can handle).

James suggested using text only as the reference to MathML. In this way, we do not change the SMIL portion of the spec, which would require a re ballot. Jonathan felt using text was a hack because it was being used for different things in different situations. James agreed, but felt that the Zed committee may need to use hacks from time to time to handle new situations.