Distressingly little comment on mode/modeless difference. There are reasons people want to be in a modeless editor, and there are reasons people want a distinction between insert mode and other modes. The primary drive for vi and emacs was the emergence of full screen editing? Sure. That was true, I lived across this window.<p>But.. it was possible to choose to get a 'page' worth of view of your teco or ed or SOS edit session, and return to a sense of being mid-page, if you wanted. It took some finicky commands but once you decided you needed a screen of context there was a way to do it, and not lose your point in a line.
(certainly, not lose the line you are "on")<p>For a DECwriter, this made very little sense to use frequently. It was best not to have to force it to print a lot of paper. For a slow ADM3/5 or VT100 it made some sense, because the repaint speed was a major pain.<p>Once you had sufficient responsiveness (2400? I would say that was about it. 4800-> was a no-brainer) then the repaint delay was low enough it was worth the investment.<p>There was another universe: Half-duplex. We just pretended it didn't exist, and that local mode entry didn't exist and designed systems not to even think about it. Odd, when you think that X10 was primarily about converting from server side to client side burden, which <i>is</i> local mode.<p>(server/client being reversed concepts in X10/X11 its confusing to talk about this, local being the key point)<p>I think vi took off, because a substantial proportion of the world did not want to think programmatically in LISP and wanted the reassurance of a moded editor, to do things. Emacs took off for people who wanted to hack the editor, more than the limited vi macro facility, and who wanted to be "in" the editor all the time. VI users wanted to distinguish being "in" the body of text/code and being in 'move around, and think about it' states of mind.<p>Had I turned left in 1979, I would have walked into a teco training room, and naturally been Emacs ever since. I walked right into a TOPS-10 SOS world, and emerged through PIP into ed and then ex/vi and then vim in the late 201x era. I personally preferred Keith Bostic nvi for a long time, and I still miss some of its innate simplicity. I was a late arrival in colourised editors. (this is not an analogy btw: it was literally random choice which room of two you went into in the first Year CS degree training for computing. We all used punched cards before this, and it was a DECwriter lab with very preciously guarded pre-ANSI terminals, too few to go around. Some of the DECwriters were converted to greek for APL, and we had a tectronix vector display which could be used as a text display device too, if you could stand the burn-in)<p>What often gets missed, is MIT, X and the emergence of EMACS key bindings as the go-to for line edit in the web. This and bash, had huge influence<p>So we now have a world of VIM users using classic VI bindings in editor, but using EMACS to do all line edit on the shell an d in the omnibox, all because of MIT and the X windows system and how that influenced graphical display norms when the first browsers emerged.