Skip to content

NVDA + MathCAT reads math elements twice in a PDF produced from LaTeX using math/setup={mathml-AF, mathml-SE} #521

@pseeburger

Description

@pseeburger

When I compile my Overleaf latex document to a PDF/UA-2 using the following preamble with math/setup={mathml-AF,mathml-SE}, NVDA+MathCAT reads all math expressions two times when opened in FireFox.

\DocumentMetadata{ lang=en-US, pdfstandard=ua-2, pdfversion=2.0, testphase={phase-III,math}, tagging-setup = {math/setup={mathml-AF,mathml-SE}, extra-modules={verbatim-mo}, table/header-rows=1} }

Is there a way to avoid this double reading, while keeping both versions of mathml in the PDF?

The mathml-SE is needed by Adobe Acrobat when NVDA+MathCAT are used to read the equations in the PDF. It does not read the math correctly at all with only mathml-AF in place.

But in FireFox, the mathml-AF produces cleaner, clearer reading of the math, e.g., a piecewise function using the cases structure.

Is there a setting in MathCAT or in my latex compilation that would set a priority between the two versions and avoid this repetition?

If not, it seems that it may be helpful for MathCAT to automatically choose to read the mathml-AF version, if it were available and does not cause an error, and not read both versions. It could then fallback to the mathml-SE version, if the AF version does not work. There could be a setting to actually still read both versions (for testing purposes), but I can't imagine most people would find it helpful to hear the math read twice.

Metadata

Metadata

Assignees

No one assigned

    Labels

    questionFurther information is requested

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions