I like articles that discuss out of ordinary approaches. This is one of those, and I relate to the pain the writer also has.
However...
> Different people use windows totally differently.
The older I get, the more impatient I am with this. In the industry, we normalized that high paid software developers (who are way better programmers, or infra people than I am) waste hundreds of engineering hours with their absolutely broken window management, and command-tab hell. Whenever I have to witness a screenshare of a colleague, my time is also wasted.
I wish this would be an interview gating: show us how you manage yourself in this big monorepo, switching tasks, windows, tabs, desktops, SSHs etc. Do not normalize noob time wasting.
Different people also use pianos differently. Yet, we don't appreciate a noob pushing random keys around it. It would be unpleasant to the ear, just how unpleasant (painful) it is to witness macOSs/gnomes bubblegum horror of sloppily animated chaos.
Back to my point: windows (as in viewports) are with us for many decades. I don't care if you tile, stack, infinite scroll them. Just be productive. Invest upfront in learning, be fast, and be able to explain what I you try to show me on screen.
There will be no one size fits for all, and that part I understand. But please have more dexterity with your desktop abilities than my grandma.
This is false analogue. Comparing SWE to pianist and using desktop as the similar element is false. Desktop to SWE is like … how the concert pianist arrives at the concert venue. You want the pianist to arrive on time but you really don’t care about their mode of transportation.
Besides it’s good to take things slow when face to face with people.
Or how a carpenter arranges their tools.
Honestly, I'd rather work with engineers trying to build maintainable code than engineers who spend all the time trying to be a power user in the command line. It's possible to be both, but bad code wastes my time much more than a demo with a couple of hiccups.
I didn't mention cli on purpose, and I don't care if they can rearrange their commits in the command line or in some fancy gui, as long as they can do it. Also rarely care about aesthetics (e.g. ricing).
I'd rather observed correlation of experts who can navigate their GUIs / TUIs, and being good software engineers.
It's hard to argue with this, yet, I never learned to touch type or type fast, never bothered to learn a bunch of vim or emacs hot keys or rice a vim setup. I use pretty basic editors and ides that barely qualify as ides.
Everyone else around me blazes at the simple act of typing. And far fewer but some have powerful vim setups where they can do inarguably useful things fast.
You would be agonized to watch me type code.
But for 30+ years it never mattered. I spend 10x more time thinking than typing, or jumping around refactoring 300 things all over the place every 5 minutes, and propduce results that work, are efficient, and fail far more gracefully than everyone else.
(I'm not claiming to trade iteration for get-it-right-the-first-time, I obviously iterate and I'd even say the correctness of the final result is mostly the result of iteration).
Maybe it doesn't conflict with your point because even so I wouldn't exactly say I'm no more functional than your grandma, so maybe typing and vim/ide-ricing are just a couple of exceptions that I can afford to get wrong as long as everything else is good.
Though I don't know what else there even is to be "fast" at. I don't think I really even understand how one can be good or bad at mere window management. ssh-ing? What is there to that? knowing how to manage keys?
However, all in all, I like this. Even if it means I'd currently fall on the wrong side of this idea, I think I agree with this idea to have an expectation of basic competency with the tools in any field one would presume to get paid for. Maybe even now I should just suck it up and learn to type and develop a powerful vim setup.
GP reads to me as “when sharing your process, be fluid instead of bumbling.”
Things like “uh, where’s that file again?”, “the window disappeared”, etc shouldn’t happen during screen sharing, pairing, and demos.
In other words: be a good presenter.
Oh god if that's all, yeah ok the bar is far lower than I was thinking. But yeah I've seen that so yeah.
Yes, the bar sometimes I have to witness, is low, very very low. I say that someone like you who even thinks and can rephrase thoughts about this, is already of the upper percentiles of capability.
OTOH presenting is a different skill. Maybe I should have elaborated that as a field expert of a (not-so-)niche thing, I often get contacted by random colleagues who will ask me to "let them quickly share their screen to show what the problem is". Their demonstrated horrific inability to "do computers" is surprising and upsetting.
Therefore I thought about this a lot in the past. Maybe the windows concept is wrong? Should there be more experiments like OP proposing? And slowly I start to believe, that probably the windows concept is okay, since you and me and thousands of others are capable, and it's just a skill issue.