I do not see any questions here about how to convert APA (or MLA, Chicago, Turabian, etc.) style into HTML (edit: styled with CSS) for online publication. Since HTML is perhaps the most accessible format for assistive technologies, I'm puzzled. As a professional who helped faculty create and teach online courses, I have experience with both website design and academia. One problem I've wrestled with has been APA run-in headings (level 4 primarily). I finally submitted the problem to Stack Overflow, and a CSS (cascading stylesheet) guru wrote a brilliant solution. My primary question is this: has anyone here encountered this problem? I'm also wondering, is the crossover between academia and online technology insufficient to know this is a problem? Is anyone here aware of issues in converting academic writing into accessible web-based material? Has anyone asked a webmaster to convert a paper into a web document and been told it was too difficult?
No, have not encountered this problem: some publishers do aim for an online presentation of materials using regular web pages indeed, but then they have a freedom to pick a convenient citation style and whatnot. And if you are not a publisher, the "normal" way to disseminate the research results in academia is a publication list on a personal page and PDF files. So in a nutshell, there is not enough interest to the specific problem you are posing here indeed.
APA and other conventional scientific writing style guides are not well-suited for online publications.
The most elaborate way of displaying scholarly articles for online viewing is probably the one used by Research Ideas & Outcomes at Pensoft Publishers.
Browsing through the XML, you will see how standardized it is and how many metadata it embeds -- here is a screenshot of one part of the
<article-meta>-tag (containing information about the submission history, license, abstract, keywords, figure counts etc.):
Such a metadata-rich formatting allows them to track statistics such as which Figures were viewed or downloaded most often - - again an example from the paper above:
EDIT: Just found out about DocMaps, this might be relevant here too: https://docmaps.knowledgefutures.org/