Re: Keyboard usage on some Gnome windows not working

[Shaun McCance <shaunm gnome org>]

On Fri, 2005-10-21 at 13:14 -0200, Matthew Thomas wrote:
> On Thu, 2005-10-20 at 09:47 -0500, Shaun McCance wrote: 
> > On Thu, 2005-10-20 at 08:49 -0200, Matthew Thomas wrote:
> > > On 19 Oct, 2005, at 2:49 PM, Bill Haneman wrote:
> > > >
> > > > Matthew said:
> > > >>
> > > >> In Windows 2000 and (I think) Windows XP, all access key underlines  
> > > >> are hidden by default.  
> > > >> <http://www.microsoft.com/enable/training/windowsxp/  
> > > >> hideunderlines.aspx> This makes the interface less ugly, and possibly  
> > > >> also somewhat faster for people who aren't disabled (as it  
> > > >> discourages them from thinking that finding and typing the access key  
> > > >> is faster than using Tab or the mouse).
> > > >
> > > > I am hoping there are missing <sarcasm> tags around that last comment  
> > > > :-)
> > > > ...
> > > 
> > > No, there aren't. <http://asktog.com/TOI/toi06KeyboardVMouse1.html>
> > > 
> > > I see there is some research showing that the keyboard is faster for  
> > > common commands  
> > > <http://ad-astra.ro/research/view_publication.php? 
> > > publication_id=1508&lang=en>, but that wouldn't include access keys  
> > > unless you were encountering particular dialogs or alerts very often.
> > 
> > The keyboard is clearly faster at many tasks.  For instance,
> > I type anywhere from 130 to 150 wpm, depending on the day.
> > If you tell me I could mouse that faster with GOK, I will
> > laugh at you and ignore every post you make from now on.
> 
> This thread is about keyboard equivalents to menu/toolbar/title bar
> commands. If you construct straw men like "the mouse is faster than
> typing" when no-one suggested that, I will laugh at you but keep reading
> your posts. I like laughing.

>From Tog's article:

* Test subjects consistently report that keyboarding is faster than
mousing.
* The stopwatch consistently proves mousing is faster than keyboarding.

There's no qualifiers here.  In the absence of any qualifying
statements,
I'm forced to assume he means "for all tasks" or at least "for most
tasks".
I'm pointing out that typing is a task, and for many people a very
common
task.  As a programmer, and as somebody who writes raw DocBook by hand,
I spend the majority of my time typing.

This article, like most stuff I've read from Tog, is sloppy.  It takes
some statistical data that's probably relevant to a lot of things we
might discuss here, and wraps it in universal dogma.  This, in turn,
creates legions of OSNews posters who think they understand the very
essence of existence because they read an "Ask Tog" article.

> > If I'm typing something in a text editor, and I want to
> > save my work, I hit Ctrl+S.  This is a very easy and very
> > common shortcut, and it's ingrained in most people's muscle
> > memory.
> 
> Actually, it's not. Lane, Napier, Peres, and S�or (2005) found that
> when saving in in Microsoft Word, people reported using the toolbar icon
> about 50 percent of the time, and using the menu item about 35 percent
> of the time. Ctrl+S was used only about 7 percent of the time.
> <http://www.owlnet.rice.edu/~peres/Rice/IJHC1802_Lane.pdf>
> (Unfortunately that is a survey, rather than direct observation; but if
> it's biased at all I would expect people to claim they use keyboard
> shortcuts more, not less, than they actually do.)

Interesting.  So my claim of "most" was off.  It was a bad word
choice anyway.  I would argue that 7% is still a statistically
significant percentage.  There are ~297 billion people living in
the US.  7% of that is ~20 billion.  That's entirely too many
people to ignore.

(Yeah, yeah, they don't all use computers.  But it's still a big
number, especially when you expand it to the computer-using world
population.)

> > Furthermore, since I'm already typing something,
> > my fingers are already located on the keyboard, very likely
> > placed nicely along home row.  Context switches can be very
> > disruptive.
> 
> Amusingly, you wrote that message in Evolution, in which Ctrl+S in the
> composition window saves your message, but Ctrl+S does in the main
> window exactly what Ctrl+F does in the composition window, and Ctrl+F in
> the main window does something else again. Have you reported this
> unnecessary disruption as a bug?

No, I never noticed it.  Did you file it?  I never noticed it
because it's never entered into my work flow.  I don't ever
save messages in Evolution.  I'm sure some people do, but I
don't.  I write them, and then I send them.  And I can't be
quite sure, but I don't think I've ever done a Ctrl+F style
search.  I search for messages all the time, but emails tend
to be short enough that I don't feel a need to search within
them.

> > Access keys are surely not as simple as shortcut keys, and
> > it wouldn't surprise me if they're slower when you have to
> > scan things while using them.  "I want File, that's Alt+F.
> > And now Save As, that's A."  But once memorized, Alt+F A
> > is fairly easy, even though the Alt key is typically in an
> > awkward position on keyboards.
> 
> Again, in Evolution Alt+F A works in the composition window, but not in
> the main window -- and this time there isn't even the excuse of Alt+F A
> in the main window doing something else.

Eh, I grabbed the first access key in front of me that didn't
also have a shortcut key as well.  I use drag and drop pretty
much exclusively for attachments.

> > If my hands are already on the keyboard, along home row,
> > I can hit Alt+C (Cancel on a lot of dialogs) really, really
> > fast.  Like, before-you-can-blink fast.
> 
> That's an unfortunate example, because having to swat away dialogs in
> the first place is a symptom of a design bug elsewhere. (For example,
> the Print Screen key is so close to the Backspace key on many keyboards
> that the "Take Screenshot" keyboard equivalent should include a modifier
> key.)

Yeah, nagging dialogs are evil.  But they're still a reality,
at least for now.  Alt+C is also the common accelerator for
Close, which you have on most preferences dialogs.  And those
aren't nagging dialogs.  They're things you asked to see.

As for Print Screen, I think there's only so much we should
be expected to do in software to accommodate the fact that
people make really crappy keyboards.  I swear, the whole
world just forgot how to type in the last ten years. 

> > When Tog says "The stopwatch consistently proves mousing is
> > faster than keyboarding." I'm curious what he means.  I'd
> > really like to see what tests were performed and what the
> > actual results were, rather than a one-line synopsis.  As
> > a mathematician, I'm suspicious of pretty much any one-line
> > statistical synopsis.
> 
> Agreed, more detailed research would be nice. Lane et al. 2005 found
> keyboard equivalents to be faster for Copy, Open, Paste, and Save, but
> their experiment seems rather unrealistic -- people were given several
> practice runs first, allowing them to keep the keyboard equivalent in
> their short-term memory, when in real life it wouldn't be. (It's the
> retrieval from long-term memory which is the big time-sucker, according
> to Tognazzini.)

Well, I think it's fair that you test shortcuts when they are
ingrained in muscle memory, because that's when they're really
useful.  If you have to stop and think about the shortcut, it's
not really useful.  If you only test shortcut keys on people
who don't have the finger habit, you're not really conducting
a very fair test.

> >   Imagine this test:
> > 
> > "A group of average users with varying keyboard skills were
> > asked to do tasks related to managing files and folders in
> > a graphical file manager.  They were first asked to perform
> > these tasks using ONLY the keyboard, then ONLY the mouse."
> 
> Yes, that would be a more realistic experiment.

But not a very good one, I'd argue.

> > ...
> > People can use the keyboard when they think it's best and
> > use the mouse when they think it's best.  And that's not
> > necessarily tied completely to the task.  It could just
> > be a matter of where their hands are.
> > ...
> 
> What do you mean by "best"? If you mean "quickest", then that doesn't
> seem to be true. Peres, Tamberello et al. (2004) found that on average,
> people who currently don't use keyboard shortcuts *strongly disagreed*
> with the statement "I would start using keyboard shortcuts if I thought
> they would save me more time".
> <http://chil.rice.edu/research/pdf/PeresEtal-HFES.pdf>

By "best", I mean "least disruptive".

Other than not having home-row-accessible cursor movement
key commands, our interfaces right now work pretty well for
both the keyboard and the mouse, as well as many other input
devices.  Why change what's not broken?

--
Shaun