Use of Machine Languages in the alt Attribute

Issue Statement

Use of TeX, LaTeX and other machine languages in the alt attribute.

Is this a practice to be condoned -- and included in WCAG 2.0 -- or does it constitute an abuse of the alt attribute? Are TeX and LaTeX special-case exemptions to what otherwise should be a rule: no machine language syntax allowed in the alt attribute? If no machine language is allowed in alt, should the WAI's advice to authors to be to add the machine language alternative through the object's own meta-data? Should the use of TeX and LaTeX, as well as other machine language syntax be included amongst the content-negotiable options for information contained in a media-specific element?

The Argument Against the Use of Machine Language Syntax in @alt

If an author includes machine language syntax in the value of the alt attribute, an author could also use proprietary and application-specific machine languages, such as Quark XPress's plaintext tags or PostScript for an image, and then argue that a "text equivalent" has been provided, even though it may be expressed in a machine language which is not ubiquitously supported by either "mainstream" applications or assistive technology.

The Argument for the Use of Machine Language Syntax in @alt

Machine Language provides a means of conveying meaningful information in the case of specialized knowledge domain markup, such as CellML, MusicXML, NewsML, the MicroArray and Gene Expressions (MAGE) or the Minimal Information Requested in the Annotation of Biochemical Models (MIRIAM).

The Argument for an Exception for TeX and LaTeX

TeX and LaTeX are capable of providing a "true" mathematical expression to an assistive technology, as described in ARIA 1.0 for aria-role="math" -- which, as of the 5 August 2008 Editor's Draft, read:

Role: math

Content that represents a mathematical expression.

This is for sections that represent math, such as images and ASCII art, but are not in a formal mathematical language. Such images MUST be labeled by text that can be converted to an accessible format, using the describedby property with a reference to a description of the math expression as it would be spoken. This is designed to facilitate conversion to speech. The text description should have no special markup used to control a speech device. Authors MAY store image alternative text in the image formats themselves. In these scenarios the alternative text from the image SHOULD also be brought into the text of the document and referenced by the describedby property.

Further Considerations

if TeX and LaTeX are special exceptions, should the WAI advise authors be to embed the machine language syntax in the digital object itself, as meta-data, rather than as the value for alt?
Assertions to be Tested:
- TeX and LaTeX are likely only to be used in specialized circumstances;
- TeX and LaTeX are likely to be encountered by students of specialized knowledge domains, who already require TeX and LaTeX aware assistive technology, or a third-party aplication-specific plug-in; on the other hand, Wikimedia is capable of accurately expressing mathematical expressions through the use of LaTeX labeling; (check with NeilS of DesignScience on this -- he is the relevant expert);
- if TeX and LaTeX are allowed, does that open a back door for other machine language code to be mis-used as alt text?