Hacker News Reader

Comment by bitwize | Hacker News Reader

I think there is something to be said for "muscle memory". Tog was testing on relatively inexperienced users. Expert Vim and Emacs users do not take two seconds to decide which key to press, and we have real corpuses of computer interaction data which couldn't exist if the two-second figure was a fast rule -- specifically, StarCraft game logs. A lot of mousing happens in StarCraft, but a lot of keyboarding happens too, and lots of productive APM just wouldn't be possible if Tog's observations were a fast rule for all users.

View on HN

> Expert Vim and Emacs users do not take two seconds to decide which key to press

Many times when editing in emacs I just "do the thing" without consciously knowing what keys I pressed.

If given a keyboard with no keys labeled, I would not easily know what each key was. Just the same, I could look away from it and type the alphabet. Magic.

I use MessageEase keyboard on my phone and you can disable the letters and just have blank buttons. I love people's reactions when they see me thumb pecking at this inscrutable black grid, it does feel like magic.

MessagEase is pretty much the only viable option for typing on my phone for me - Although I found it while I was on the search for a chording keyboard, which then didn't really meet my expectations...

(Not on a touchscreen at least)

> Many times when editing in emacs I just "do the thing" without consciously knowing what keys I pressed.

More than once I've tried to describe a feature to some other Emacs user, and had to physically put my hands on a keyboard and watch them in order to say which keys were used to invoke it.

I'd add that the opposite can be true, too. My 'command menu' is like VSCode's: you can narrow down the list of commands, and you can see the keyboard shortcut for running the command beside that.

I was given a mechanical keyboard with a set of blank keycaps (it was in my wish list). People thought it was black magic when I typed, but I never even noticed.

It's exactly like how you totally forget PINs, phone numbers, or passwords once you've typed them enough. You need a keypad in front of you (or at least to pretend there's one) and to just let your fingers do what they've done so many times before in order to figure it out.

I don't do well with laptops and mac keyboards due to their nonstandard key arrangements though.

> It's exactly like how you totally forget PINs, phone numbers, or passwords once you've typed them enough.

That happened to me once, the electrician had to change the electricity meter, he called me to ask for the code to enter the flat and I couldn't recall it; I had to go down and open the door.

I have trouble with the newer "flat key" mac keyboards because my fingers don't have curvature to tell when they're centered in the right place.

My favorite laptop keyboard is the first gen macbook pro. Great comfortable concave keys with good throw. Some thinkpads were good too.

He seems to have been open to the idea:

> I would also like somebody—anybody—to show up with one properly-conducted test report that shows that command keys are equivalent or superior to pull-down menus somewhere, somehow.

It's been 30 years since he wrote that. Has anyone in that time been able to produce data that refutes Apple's findings?

Most likely, but I don't know how to do a search for it.

The more items in your menu the slower it is to find things in it. I can type NY faster than I can find New York in your states menu (I picked a state near the middle to ensure you have to scroll, but can't scroll to the end which might be faster if you know).

Tog was forced to back off his claim, but this wasn't publicized as much. A well trained expert with key shortcuts used all the time is faster than someone with a mouse - but if you don't use the shortcut often, or there are too many keystrokes the mouse wins.

Maybe not the same, but it is a fundamental user experience principle.

"Showing users things they can recognize improves usability over needing to recall items from scratch because the extra context helps users retrieve information from memory."

https://www.nngroup.com/articles/recognition-and-recall/

Doesn't this depend on the number of actions available at a given point in time? For example, if I'm "using my Macbook", I could be wanting to do any of no less than a million (not an exaggeration) actions at any given moment. It's infinitely easier for me to open Alfred (or the OSX spotlight for a weaker example) with one keystroke and type the thing I need to do. It simply is not possible to give me a menu (presciently) for this kind of thing.

Perhaps if you're restricting context to one app that doesn't have too much going on it's a different story, but compared to 30 years ago there is a lot more people (especially devs, or Starcraft players) need to be doing on the computer at a given time.

I suspect that graphical/menu input is preferable only for a Goldilocks range of decision tree complexity. Too many options and displaying them all to the user at once is impossible or unhelpful, forcing you to provide a search function (see the Windows Start Menu). Too few options, and the user won't need any prompting to know what choices they have and what keys they can be selected with (eg. Escape vs. Enter).

Exactly. People that use old green screen systems have memorized the keystrokes. Menu 2, 7, 12 is this. 6, 10, 3 is this. Tab tab option P is print.

People use/used these systems faster than the computer could even run. The input would be buffered so all of it happens correctly but the screen would struggle to keep up.

Sometimes with fatal results. https://en.m.wikipedia.org/wiki/Therac-25

> It's been 30 years since he wrote that. Has anyone in that time been able to produce data that refutes Apple's findings?

Yes, all of the specific experiments Tog describes are bogus. In versions of the experiments where the keyboard user isn't crippled by weird limitations that don't apply to actual computer users, the keyboard is often superior to the mouse (not always, just often): https://danluu.com/keyboard-v-mouse/

Exactly this. His test methodology is BS.

Everyone ever using a POC application in a store - the DOS/Windows kind, from before the current generation browser-based nonsense?

If you saw them, you probably noted they were fully keyboard-operated, had a lot of functionality accessible under function keys, and enabled the clerks to accomplish a lot of tasks without actually looking at the screen(!). You can't beat something like a muscle-remembered "F2, enter, enter, <type name>, tab, enter, enter" with a dropdown menu. I've seen cashiers do the former faster than it'd take to select a single option from a menu.

The only way I can imagine Tog's assertion to make sense is if we're comparing a mouse menu vs. looking up a shortcut on a physical, printed cheat sheet (which is essentially a menu-equivalent), for an operation that you're using for the first time in your life. None of that resembles actual use of real software.

Starcraft supports the theory that mouse is much better for positioning and selection than keyboard. It's so much better, it's worth it keeping one hand at the mouse.

I mean, that was what the vi fans I met in my youth tried to deny. They claimed positioning by incremental search was better.

In a late uncle's stuff, I came across a copy of "The psychology of human computer interaction" by Stuart K. Card et al. A book from the early eighties detailing Xerox PARC UI research. They did a study comparing cursor keys (or vi keys), word and line keys (home, end, next word etc.), incremental search and mousing. They found the same as Apple later would.

I wonder how many of the people who did the PARC research, and the people who did the later Apple research, were the same people.

Also, the test he designed, and his methodology, was BS. Moreover, 0 peer-reviewed results.

I think "mouse is better" was a conclusion in search of evidence. It was Apple in the 80s -- Steve Jobs was a huge believer in mouse-based UIs, but most users and developers outside Cupertino were not. Thry had to be convinced that mousing is universally better than keyboarding.

Reply on news.ycombinator.com