I believe you're describing defacto habits ingrained into everyone from
years of using Microsoft's original UI design than a qualitative critique
of this particular design aspect.  Yes, familiarity can matter (see below),
but the concept of *"Point to what you want to interact with,"* is such a
natural way of thinking and working for a human user.

Having keyboard focus decoupled and *way off* somewhere else than where the
users' attention has since moved, often leads users to a similar confusion
that you described, but just from the other direction -- the user wondering
why *nothing is happening* when typing into a field -- but this time
because the list in which they last clicked just to check something
real-quick still has the keyboard focus.  This is basically what happened
to Eliot..

Coupling it to the mouse pointer helps remove all confusion about where
keyboard focus is, and is a natural concept that can be assimilated in
about a day.  Dump Windows for Linux, and the behavior can be extended into
the host system, too.

Now, maybe for a single-use Form with a few fields used by external users
for something simple, designing it click-for-focus may be worth aligning to
Microsoft's defacto.  But for all-day developer work involving multiple,
multi-paned windows, allocating power to pointing makes the system feel
"light and responsive" instead of heavy and cumbersome.  For me, anyway...
:)

Best,
  Chris



On Mon, Sep 21, 2020 at 8:12 AM Ron Teitelbaum <ron at usmedrec.com> wrote:
>
>
> On Sun, Sep 20, 2020 at 5:14 PM Chris Muller <asqueaker at gmail.com> wrote:
>>>
>>> >  For example, imagine the power of having EVERY PIXEL on your desktop
being ready for ANY INPUT, regardless of window z-ordering, simply by
moving pointer there and pressing a mouse or keyboard button.
>>> >
>>>
>>> Here you see, we differ. It's likely partly due to growing up using
RISC OS and the original ST-80 UI ...
>>
>>
>> I grew up using MS Windows, the beast that made "click for focus" the
defacto disaster it is today.  Not everyone is willing to change how they
work, but a few years ago I got a tendonitis that really made me notice
that I was doing, literally, hundreds upon hundreds, of extra clicks every
day for no other purpose than reassigning keyboard focus, which, as you
know, have to even be "strategic clicks" (window title bars, borders, etc.)
to avoid suffering unwanted selection change, otherwise having to click yet
again just to "get back" to what I as looking at.  Not only the pain, but I
realized what a total time sink it was.  The only way to escape was I had
to be willing to change, which actually turned out a lot easier than I
thought...
>>
>>
> Hover to select is one of the greatest causes of user confusion that I
encounter.  Most windows users are used to clicking and moving the cursor
out of the way so they can see.
>
> Now imagine a text box.  As a user you click on it so that you can type
into it.  You move your mouse away to type into the box BUT NOTHING
HAPPENS.  Why because your moving the mouse away from the text field took
focus away from the text field.
>
> Or you click and start typing but your mouse pointer is in the way so you
move it and now nothing you type goes into the field.  We worked around
this issue by locking some of the important fields when a user selects it
but not all of them so it's not consistent.
>
> We also have the issue that if your mouse happens to be hovering above a
button things work for a minute but then stop working because the button is
capturing keystrokes.  The user only moved their mouse. They didn't mean to
hover over a button or they clicked a button and got the result they
expected but had no reason to move their mouse cursor off the button
afterwards.
>
> While I understand the unnecessary click argument and the harm that
clicking does to your wrist, there is still a lot to be said for
consistency, user understanding, and user experience.
>
> Just my 2c.
>
> Ron Teitelbaum
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20200921/7e0ea9dc/attachment.html>