keyboard accelerators

keyboard accelerators

[David]

> >keyboard accelerators are indicative of a broken user interface.

Boyd is, as usual, right.

> Not everybody is coordinated enough to use a mouse efficiently.

Ditto keyboards.

Let's not be restrictive:
many of the people here use trackballs instead of mice,
because they don't have the desk space to conveniently move a mouse around.

The input *device* isn't under discussion here:
it's what you expect it to do with it:
either to indicate geometric information,
or to input things which are commonly interpreted as characters,
but might be used for other event-like stuff.

> My mother has difficulty, for example.  Even worse, some people
> have genuine physical handicaps who *cannot* use a mouse, but who
> can use some form of keyboard.  Or vice versa, for that matter.

This says more about the inadequacy of input devices than about input
paradigms.

Even though I'm, fairly able bodied,
I've wanted one of those neat eye movement things,
so I can just look where I want to,
instead of having to move the mose/spin the ball/waggle the joystick/whatever.

> Recently my mouse broke.  It was a weekend.  Under X, I was completely
> paralyzed.  I rebooted my PC to Windows (blech!) and was able to get
> some useful work done in a multi-window environment.  It was painful,
> but better than the complete paralysis under X.

So you window manager sucks:
what has that got to do with the price of mice?

To be fair to X (not that it deserves it),
there's nothing that I am aware of that forces you to use a mouse:
some window managers (olvwm, for one),
can (theoretically) have a mouseless interface,
parts of which I occasionally use (when I'm forced to use a trackball).

> (Argh, mice are expensive.)

(And they nearly all suck!)

> Supporting multiple styles of input strikes me as generally useful.

Well I'm afraid I disagree:
supporting multiple input *devices* may be useful.
but if you have a good input style, why clutter things with an inferior one?

> I don't think it necessarily implies software bloat.

(All together now children):
 OH YES IT DOES!!!!

	Dave.

keyboard accelerators

[Castor]

> forget typeahead.  multitasking is a sign of real bloat.  quote from
> the New York Times, 25 April 1989:
> 
>   Real concurrency---in which one program actually continues to
>   function while you call up and use another---is more amazing but of
>   small use to the average person.  How many programs do you have that
>   take more than a few seconds to perform any task?

The New York Times is not the only authority to have said this.

In Nikalus Wirth's Project Oberon one of the great simplifications
he makes is to not have multitasking. Instead concentrates on making
the programming environment so dynamic that one doesn't miss it.

Obviously there are tasks where this model is simply inapplicable,
but if you can rebuild your kernel in a few seconds. . . 

	-castor

keyboard accelerators

[Felix]

> I think like Chuck Moore that 'typeahead' is another sign of bloated
> software. Nowadays i am relly disapointed if my 50+ mips workstation
> is not keeping up with my typing.

forget typeahead.  multitasking is a sign of real bloat.  quote from
the New York Times, 25 April 1989:

  Real concurrency---in which one program actually continues to
  function while you call up and use another---is more amazing but of
  small use to the average person.  How many programs do you have that
  take more than a few seconds to perform any task?
--

keyboard accelerators

[serge]

>I think like Chuck Moore that 'typeahead' is another sign of bloated
>software. Nowadays i am relly disapointed if my 50+ mips workstation
>is not keeping up with my typing.

	% make	# er, excuse me, mk :-)
	# switch to other window to answer 9fans mail in the meantime :-)

keyboard accelerators

[rob]

	PS: How good is an plan 9 terminal with ISDN to another CPU server really?
	    (How fast is graphic output? Or is that only possible with ASCII?)

For interactive 2-D stuff, it's fine: it keeps up with the input for interactive
programs, menus, and so on.  It feels faster and more smoothly responsive
than applications running under X windows.  But I wouldn't recommend it for
texture-mapped rampaging dinosaurs.

-rob

keyboard accelerators

[Heiko]

> 	+ you can't do mouse ahead the same way you can do typeahead,
> 	  e.g. the mouse (pointer) can be in an arbitrary position, so
I think like Chuck Moore that 'typeahead' is another sign of bloated
software. Nowadays i am relly disapointed if my 50+ mips workstation
is not keeping up with my typing.

	Heiko Wengler

PS: How good is an plan 9 terminal with ISDN to another CPU server really?
    (How fast is graphic output? Or is that only possible with ASCII?)

keyboard accelerators

[Scott]

| + anyone remember piewm (X based window manager with
|   circular instead of drop down menus)?

Yes, Don Hopkins' pie menus were really cool.  Originally done for
NeWS, he also used them in the Unix port of SimCity using Tk.

keyboard accelerators

[serge]

(Gee, the fire seems to be dying down.  Now, where did I put that
gasoline?  Oh, yeah ... WHOSH! :-)

Some random thoughts:

	+ Apple, which was mouse based from day one (TM) (R) (C) (:-),
	  has keyboard accelerators ...

	+ so does Windows (although you could say that it's because they
	  were trying to introduce mice into predominantly keyboard
	  based environment; also, Windows95 appears to have [much?]
	  less keyboard accelerators); still ... (both they and Apple
	  must have done some user interface studies; mustn't they?  :-)

	+ you can't do mouse ahead the same way you can do typeahead,
	  e.g. the mouse (pointer) can be in an arbitrary position, so
	  the motions needed to get it to the ``menu button'' (which can
	  also be in an arbitrary position, and move as well) differ
	  each time, unlike the keyboard (e.g. Ctrl+Esc,p,Return to get
	  to the Program Manager in W ... er, but I digress :-)

		+ anyone remember piewm (X based window manager with
		  circular instead of drop down menus)?  it was trying
		  to address some of these problems, e.g. you could
		  memory train your hand much easier: just click, pull
		  the mouse to the east and release

	+ we are still talking about predominantly keyboard based input
	  methods (until voice recognition becomes commonplace :-), so
	  most of the time you are typing, rather than clicking with the
	  mouse; even Pike (who had to quit his Help paper because he was
	  about to have to start using the keyboard :-)

keyboard accelerators

[Nigel]

Nice summary

keyboard accelerators

[Felix]

> > keyboard accelerators are indicative of a broken user interface.
> sure is nice to get such compact bits of wisdom and enlightenment
> in my mailbox once in a while. i will meditate immediately.

om mani padme hum.  the premise is that mice are better than
keyboards.  the evidence is time-motion studies.  therefore, if an
interface requires keyboard accelerators, then the mouse interface is
flawed.

well almost.  but anyway, it's useful to aggressively push this
viewpoint, because mice really are poorly utilized when people layer
them onto a traditional keyboard interface.  even when the interface
is built from scratch, people often ignore usability issues in favor
of the eye-candy school of design.

(well almost.  eye-candy is often useful.)
--

keyboard accelerators

[Mike]

>keyboard accelerators are indicative of a broken user interface.

Not everybody is coordinated enough to use a mouse efficiently.
My mother has difficulty, for example.  Even worse, some people
have genuine physical handicaps who *cannot* use a mouse, but who
can use some form of keyboard.  Or vice versa, for that matter.

Recently my mouse broke.  It was a weekend.  Under X, I was completely
paralyzed.  I rebooted my PC to Windows (blech!) and was able to get
some useful work done in a multi-window environment.  It was painful,
but better than the complete paralysis under X.

(Argh, mice are expensive.)

Supporting multiple styles of input strikes me as generally useful.
I don't think it necessarily implies software bloat.

keyboard accelerators

[ozan]

> keyboard accelerators are indicative of a broken user interface.
sure is nice to get such compact bits of wisdom and enlightenment
in my mailbox once in a while. i will meditate immediately.

ommm.

keyboard accelerators

[Boyd]

keyboard accelerators are indicative of a broken user interface.

keyboard accelerators

[forsyth]

>>By keyboard accelerators, I assume you mean keyboard shortcuts for common
>>operations, such as "go to end of line" or "select the next word". I have
>>used such shortcuts so heavily in MSWord that I have watched for them in

the context was Acme on Plan 9, where `shortcuts for common operations' are
unlikely to be as beneficial as you might think, not least because any
text on the screen is `operative': just click on it using
a suitable mouse button.  it's active text with a vengeance!
consequently, the things usually bound to keyboard accelerators
are not necessarily the `common operations' in Acme (more accurately, in its client applications).  furthermore,
your hand is probably already on the mouse, owing to the way the
interface works, and it would be more effort to shift to the keyboard!

see the Acme paper for examples.

keyboard accelerators

[Don]

 [Apologies for joining into the middle of a discussion - I just started
lurking the list 4/18, so the earliest msg I see is dated 4/11, in the
digest dated 4/19! So far, not near the traffic that DrWho gets! :-) ]

By keyboard accelerators, I assume you mean keyboard shortcuts for common
operations, such as "go to end of line" or "select the next word". I have
used such shortcuts so heavily in MSWord that I have watched for them in
any system level editor I use, and missed them sorely when absent. Using
MSW on the Mac gives me a complete WintelPC numeric keypad with lots of
neat actions. Most editors I have used only support the arrow keys, maybe
(rarely) PgUp/PgDwn.

Well-done Mac applications support the mouse, since it is totally assumed
to be present. It may have become "standard" on the Wintel systems, but
after using one for a few minutes, I begin to understand people's
scepticism about mice! In applications I use often, I use mouse or keyboard
depending on where my hand is when I want to perform a command. In those I
use rarely, I like to have the expected keyboard shortcuts (Win3 seems to
have not achieved much standardization yet, at least not within offerings
from MicroSoft) and I do get brought up short when one of the really
standard ones is not present, or is used in strange ways. I just started
Quicken, for example, and the standard print command prints checks, not the
report I just brought up on the screen, and it is not context-sensitive. I
will definitely complain about that one, if it is not fixed in Q5!

This reflects the main difference I see between Mac and Other: consistency.
When I want to quit from the UNIX mailer I use at the office, an "x" is the
normal save/quit and a "q" quits without saving. At home, with /bin/mail,
the senses are reversed. If one of these is asserted to be a standard, why
hasn't the other been fixed? (well, the system here at work WAS obsolete
when it was procured by DOD in 1987!) The extended keyboard on the Mac has
a delete-forward key which I have never had before - works great! MacOS
hardware & software have a much shorter history than others, and standards
have been rigidly enforced by the user community to an unprecedented level

Don Ellis
donls@michelob.wustl.edu

Newbie/lurker extrordinaire![sp?]

keyboard accelerators

[Nigel]

> >>If I (as a hypothetical example) add keyboard accelerators to Acme,
> 
> >150 hours of community service for that one, i think.
> 
> I'm not so sure about this.  A mouse is nice, true, but it takes time to get
> your hands from the keyboard to the mouse and back again (especially back
> again, since you have to reposition them, e.g. if you're a touch typist).
> 
> Until we have better/more integrated keyboard/mouse input devices (e.g. on
> one of the PC shows I saw a keyboard split in two, with each half acting as a
> mouse by you just moving it around with [roughly] the palm of your hand), I
> think the keyboard will remain very valuable.  For example, (one of the reasons
> :-) I don't like Netscape is that it doesn't have easy keyboard shortcuts.
> 
> What do people (especially Plan9 ones :-) think?
> 
Point 1: I use Netscape heavily, and never noticed it had 
accelerators. I think this speaks volumes.

Point 2: MSW claims to be drivable by keyboard or mouse. The reality 
is the mouse is vital, and the keyboard occasionally useful. Too many 
applications _shirk_ providing decent mouse interactions because they 
must support the keyboard as well to keep His Eminence happy.

Point 3: I find having to use the mouse to make small local movements 
in sam (eg up/down 1 line) a pain when I am in full typing flood. 
This is about the only specialised keyboard stuff I miss. I guess 
because I used predecessors of the Plan 9 WM before I used 
MSW, I've a pretty good idea that acceleration is not all it's 
cracked up to be. Also when you see, eg, all the X crud needed 
to cope, the minimalist instinct comes on very strongly... 

Summary. No, please, no, except maybe up and down arrow when editing.