Yes, Pandoc goes a good job as long as the input is constructed carefully. That is one thing I like about it for generating instructional materials.
My primary output format is HTML. The existing WAVE (
http://wave.webaim.org/ ) tool seems to work well as a post-processor. The HTML template would likely need to be tweaked to generate accessible documents and authors will need to write to certain conventions, of course. There may be other tools for other formats.
However, it might be possible to write a filter that does some checks on the Pandoc AST to warn authors that there may be accessibility issues, much like what WAVE does, but in general. Of course, it is likely not possible to check all the WCAG 2.0 AA issues at that level. And, this would not guarantee that the writers are well-behaved.
In the HTML format, I currently generate MathML but MathJax would support a broader array fo browsers.
On Thursday, November 2, 2017 at 5:38:04 PM UTC-5, Kolen Cheung wrote:
I didn’t recall there’s such a tool (I could be wrong though). Have you read through the pandoc-extras in pandoc’s wiki page on GitHub? It seems using a post-processor to check and work backward to tweak your template is the way to go. Although pandoc’s output generally is quite good, but pandoc is no linter as it is not designed to be. So I don’t imagine any tool other than post-processor would help. Even if there’s a template designed to conform to a certain requirements, what pandoc writes could break it that requirement. And then pre-processor/filter are even more difficult to do what you need, because it doesn’t see the resultant format.
Does the doc involve math? If you use MathJax, I think it has been built with accessibility in mind.