Accessible reviewing

Reviewers are not all sighted, and review tooling often assumes they are. This page describes what works today, what does not, and which workflow to choose.

Why the source is the accessible surface

Texchanges keeps every change in the document source as ordinary text:

\txreplace[author=phuc,id=R12,comment={Clarify this claim.},status=pending]{human-like behavior}{O\&M-aligned behavior}

A screen reader reads that line completely: who proposed the change, which change it is, why, whether it is still open, the old wording, and the new wording. Nothing is hidden behind a colored underline, a hover target, or a sidebar widget.

This is a real advantage over word processor track changes, where the same information lives in interface chrome that must be navigated separately from the text. Here, reviewing the source is reviewing the changes.

The practical consequence: if you use a screen reader, work in the .tex file with your usual editor, not in a rendered PDF.

A terminal-only review workflow

texchanges-merge covers the whole decision cycle without a graphical interface. Every step prints plain text.

Read what is proposed, without changing anything:

texchanges-merge paper.tex --accept --dry-run

That prints a unified diff. Decide one change at a time by its ID:

texchanges-merge paper.tex --accept --id R12 --in-place
texchanges-merge paper.tex --reject --id R13 --in-place

Decide everything from one reviewer at once:

texchanges-merge paper.tex --accept --author phuc --in-place

Or step through interactively, answering a, r, or s per change:

texchanges-merge paper.tex --interactive --in-place

--in-place writes a .bak copy first, and malformed input fails before anything is written, so a mistake is recoverable.

The change report is also available as text: compile with \txlistofchanges and read the generated .txc file directly rather than the PDF.

Low vision and color blindness

Do not rely on the default palette. Author identity is conveyed by color alone, which fails for color-blind readers.

\usepackage[review,markup=nocolor,authormarkuptext=name]{texchanges}

• markup=nocolor turns off color entirely and distinguishes changes by underline and strikethrough instead.
• authormarkuptext=name prints the author’s full name rather than a short ID.
• authormarkup=brackets or authormarkup=footnote moves attribution into the text flow instead of a small superscript.
• commentmarkup=footnote moves notes out of the margin, where they are usually set in a smaller size.

For magnification, \txsettruncatewidth controls how much excerpt text a report line shows before it is cut, which matters when lines wrap heavily at large zoom.

Current limitations

These are real, and we would rather state them than let you discover them mid-review.

• The PDF is not tagged. Texchanges does not yet emit PDF/UA structure, so assistive technology reading the rendered PDF gets no semantic marking for a change.
• Deleted text is announced as ordinary text. A screen reader reading the PDF sees struck-out words as normal words. Someone listening to a review-mode PDF may hear the old and new wording run together with no indication which is which. This is the single strongest reason to review the source rather than the PDF.
• Author color is not announced. In the default preset, author identity exists only as color in the PDF. Use authormarkuptext=name so the name is in the text.
• soul and ulem markup is visual only. The underline and strikethrough carry no semantics that a PDF reader can expose.

Tagged-PDF support, and accessible visual presets verified against real assistive technology, are tracked under “Robust and accessible authoring” on the roadmap and are part of what 1.0.0 must deliver.

Reporting problems

If a workflow here fails with your assistive technology, please open an issue and say which tool and version you use. Accessibility reports are treated as bugs, not feature requests.