> In all cases however, the program without the goto statements turned out to shorter and more lucid<p>This was with ALGOL though; Haskell uses lambdas and is shorter still. I feel like Scheme / callcc might be even shorter. And the difference between callcc and goto is not at all clear.<p>To be fair, though, he does mention the human-computer interface. Unfortunately this has still not really been explored in the context of programming. There are toy environments like Light Table etc., and the more developed languages come with IDEs, but nobody seems to use them in practice, preferring vim or emacs or whatever.<p>My guess for this is that writing GUIs in an extensible, modular, and discoverable/usable way is still an open problem. There are style guidelines, e.g. <a href="https://www.smashingmagazine.com/2016/06/designing-modular-ui-systems-via-style-guide-driven-development/" rel="nofollow">https://www.smashingmagazine.com/2016/06/designing-modular-u...</a>, but those aren't code and don't fit into an Agile-style incremental development model. The only thing that seems usable are constraint layouts, e.g. <a href="http://ieeexplore.ieee.org/document/4483218/" rel="nofollow">http://ieeexplore.ieee.org/document/4483218/</a>. But those don't account for all the little UI things, like the efficiency of the interface or which actions are most commonly used. IMO there needs to be some library to just take a bunch of actions/verbs (e.g. Copy to clipboard) and nouns (displayable data stuctures: images, documents, etc.) and turn that into an optimized user interface design, with keybindings, menus, toolbars, dialogs, widgets, etc. And then of course there's the issue that picking a GUI pretty much fixes your choice of programming language, away from any of the functional PL's.