> Blog posts are making the rounds listing all the challenges with SPAs: history, focus management, scroll restoration, Cmd/Ctrl-click, memory leaks, etc.

I guess? Like, who is implementing all this BS themselves? Every time I've written a SPA, my framework and/or routing library handled all this crap for me. None of these "challenges" were ever problems for me, and this sounds like another case of programmers creating an issue where there actually isn't one.

If you use a proper framework or routing library that changes the history state of the URL, it's pretty much impossible for users to tell whether something is a SPA or an "MPA". So who cares?