I just hope they don't stop to offer the papers in PDF. Even when I'm on a computer, I still prefer to read PDFs.
Unfortunately many institutions and businesses have ignored its limitation because PDF turned out to be an obvious-but-naive to put a 'sheets of paper' metaphor into a computer system, which in the 1990s appealed to tech illiterate folks doing bare-bones computerization of existing paper systems. So later we got complicated and error-prone tools for editing PDFs, and many random additions to the spec to allow for unusual use cases.
As an academic researcher, generally speaking I also prefer PDF, and the inflexibility and static nature is a feature, not a bug. I appreciate the fact that a paper will appear the same everywhere, that I can refer to "the top of page 7", etc.
The exception is if I wanted to just skim a paper; in this case, I think I'd prefer HTML.
I'm a huge fan of what arXiv is doing here. It effectively preserves the status quo, while adding an additional option on the side. The HTML option might prove a little bit useful for me, and it is likely to prove extremely useful for people with disabilities.
There are many great solutions to this problem, including ones that don't require Javascript at all. This website (https://gwern.net/silk-road) presents a really good example -- every header and sub-header is a clickable anchor. If more granularity is needed, on newer articles most of the paragraphs start with an italicized margin note -- though for technical writing, paragraph anchors might be better. The page also pays careful attention to print CSS and has a 'reader mode' to convert all links to footnotes when printed.
Some websites will also preserve the text you select in a URL anchor, but more often than not this is just cumbersome. It also has a greater risk of not surviving changes to the webpage.
It's actually hit ~88% of the market https://caniuse.com/mdn-html_elements_a_text_fragments but unfortunately, Firefox remains a holdout* and that's my browser, so I don't use it (although maybe I should just install https://addons.mozilla.org/en-US/firefox/addon/link-to-text-... and try it out - my existing method of making new anchors for annotation purposes is cumbersome).
* Firefox officially is positive on it but no sign of any movement on it: https://mozilla.github.io/standards-positions/#scroll-to-tex... https://github.com/mozilla/standards-positions/issues/194 https://bugzilla.mozilla.org/show_bug.cgi?id=1753933 https://wicg.github.io/scroll-to-text-fragment/