I’m happy that somebody’s finally starting to address some of the web’s big design oversights. (Span-to-span links were a feature of most pre-web hypertext systems, and certainly are and remain an important part of Xanadu.) Nevertheless, I worry about trying to tie it to a single site or platform, instead of trying to push this fix through w3c standardization.
In fact, there *is* a w3c standard for this type of link, and the HTTP protocol theoretically supports fetching specific byte-spans. Unfortunately, neither are widely implemented (by browsers in the former case, and browsers and servers in the latter), even though these standard are twenty years old. The reason is probably that such features depend upon the assumption that URLs are permanent addresses for unchanging content, and (starting with CGI but proceeding along to web apps and domain squatting) the web as a whole has totally abandoned this restriction. So, it makes more sense to use span-to-span links on IPFS or some other content-addressed protocol, wherein this restriction is actually enforced.
Just having span-to-span links doesn’t, itself, add transclusions, transpointing windows, versioning, or any of the other fundamental hypertext features that the web decided not to bother with. And, implementing span to span links in a walled garden isn’t hard — it’s an afternoon project for a beginner programmer. Hopefully, more beginner programmers will write hypertext systems, and we’ll dig ourselves out of this mess.