HTML isn't immune from broken backwards compatibility.
The frame element has been completely removed, the hgroup element is gone, and so is the dir element. acronym is deprecated in favour of abbr. isindex, plaintext, xmp, and listing are all dead.
The attributes border, clear, background and bgcolor have all been removed by HTML, and shifted to be CSS' responsibility instead.
Just moving between EPUB 2 and EPUB 3, you lose the DAISY format support, and external resources. (EPUB2 let you use full URLs to specify parts hosted externally, like webpages, but EPUB3 requires itself to be self-contained. Not a bad change, but still a breaking change.) NCX replaces just using a HTML5 nav element, and a few more things.
All of those things mean that there _are_ technical documents that exist, that _aren't_ readable without some effort to update them.
Yes, but all of those deprecated elements can still be rendered just fine by some of the current browsers. And even if, say, you couldn't display an index page with frameset, then the stuff inside the frames is just plain old html files anyway, which you can display easily. bgcolor, etc. all still render fine in current browsers as well.
The main thing is that information which is important to people will get converted to newer formats over time, just as old print books get reprinted if they're popular.
> The main thing is that information which is important to people will get converted to newer formats over time, just as old print books get reprinted if they're popular.
Vs. decent-quality old "dead tree" books, properly stored, can be ignored for centuries and still be perfectly fine.
You might want to ask a good historian or librarian about all the incredibly important (historically, to us, now) documents which we know existed, but we do not have, because there was no continuously-operated, high-budget "Holy Brothers of Document Preservation" monastery doing all the re-copy work needed to preserve them. (And maintain off-site backups in case of fire at the Monastery, and ...)
(If you aren't familiar - most of the materials used for documents in ancient times degrade fairly quickly. Unless (say) carefully stashed in a nice, dry cave in an arid climate. And even that stuff tends to be "crumble if you touch it" fragile.)
For every important historical work we have, there are several more, at least as important, that we know we're missing. And presumably even more that we've never heard of.
Neither is "LaTeX"; but, like Latex and other text-based formats, the content of the document is, in general, human-readable when it's read, unlike, for example, jpeg images.
> The frame element has been completely removed, the hgroup element is gone, and so is the dir element. acronym is deprecated in favour of abbr. isindex, plaintext, xmp, and listing are all dead.
Those are all basically cosmetic/layout related, though. The text itself is still perfectly readable.
None of that is unreadable. You can always download a browser from a few years back and open the page with that. So while there may be some inconvenience with old information, there is no loss of content.
The frame element has been completely removed, the hgroup element is gone, and so is the dir element. acronym is deprecated in favour of abbr. isindex, plaintext, xmp, and listing are all dead.
The attributes border, clear, background and bgcolor have all been removed by HTML, and shifted to be CSS' responsibility instead.
Just moving between EPUB 2 and EPUB 3, you lose the DAISY format support, and external resources. (EPUB2 let you use full URLs to specify parts hosted externally, like webpages, but EPUB3 requires itself to be self-contained. Not a bad change, but still a breaking change.) NCX replaces just using a HTML5 nav element, and a few more things.
All of those things mean that there _are_ technical documents that exist, that _aren't_ readable without some effort to update them.