Re: Justification for the button order change - the long version

[Seth Nickell <snickell stanford edu>]

> >So my biggest concern is "external to GNOME" interactions, such as with
> >KDE or Windows users. I think context here is more than sufficient for
> >there not to be a problem though, since we use a different phrasing for
> 
> People moving from a Unix workstation to a PC (or the reverse) usually
> have problems with the different keyboard layouts despite that fact
> that the context change is quite different, so I would argue that it
> will indeed be a problem.

Very funny! :-) After sending that message I went to work in a Unix lab,
and was mulling over using switching keyboard layouts as an example of
something that's pretty easy to do given proper context, and I come back
to read this ;-)

At first its non-trivial, but after switching between them for less than
a day it becomes very natural. But I may have more practice at adapting
to keyboard layouts in an abstract sense because I've had to learn to
switch between qwerty and dvorak, so I may be a very bad example case
for this.

Note, that this is a case where *muscle memory* is dominant. Its the
radical contextual similarity of keyboards that causes the problems. For
example, I had absolutely no problems with switching key positions when
I used qwerty on solaris and dvorak on PC, because my brain learned to
associate the solaris cntrl/capslock position with the qwerty layout.
Its actually more troublesome for me to type the same layout on both
systems. In this case the context is the physical feel of the keyboard,
it has little to do with the screen. So its not suprising that given
very little contextual change key changes would be problematic (also
note that they are FAR more tightly ingrained into "muscle memory" than
dialogue buttons, which won't even have a consistent spacing between
platforms or even themes or font sizes!).

But this is a pretty obscure example on my part because not so many
people use two very different keyboard layouts (even if they commonly
use two keyboards with a few keys in different positions).

> Another example, using debit cards to pay for purchases has become
> very popular.  In my area most of the debit card swipe machines put a
> picture of the card on the side of the slot where the magnetic strip
> should face.  But 2 store chains used a different machine (different
> shape and colour) that reversed that standard.  Despite knowing that
> those machines were "backwards" and that hence I had to reverse things
> when faced with those machines in those 2 stores I always swiped the
> card the "standard" (ie more common) way and had to redo it when I
> realized my mistake.

Another case where the context you give is irrelevant. Its not context!
Its just a nice background, very different things. The process of
getting the card out of your wallet, flipping it a certain way, and
sliding it through the machine is the same.

Imagine instead that it was a different machine, lets say that the card
swipe was on a different side. *THAT* would be useful context, it would
change the procedure in a way that would affect your subconscious set of
steps for swiping the card, and you would probably end up being quite
proficient at both.

Colour of the machines is a very weak context. In contrast, position of
the card swipe is an *important* context. Similarly, colour of the
window borders is a weak (or unimportant) context, but labelling of the
buttons and the dialogue itself *is* important to the context.

> >GNOME 2 settings dialog:
> >http://beauty.stanford.edu/gnome2-mouse.png
> 
> I find this example interesting because to me it demonstrates
> extremely bad UI design.  There is no way to abort any changes you
> have made (ie to return your system back to its same status before you
> starting playing with the settings).

(we're just punching all the buttons, aren't we? ;-) I'm not going to
get into a treatise on instant apply, but the basic principle is that
instant apply improves direct manipulation and a general experience of
the computer even if it loses a dubious "feature".

Various studies (one at Apple, one at Stanford, that I have heard of)
have demonstrated that with *non* instant apply users periodically used
the Cancel button (but not that often). With instant apply cancel button
use virtually disappeared, with most people choosing to change values
back to where they found them rather than Cancel. Thus dropping the
Cancel button is an obvious step.

> >That's quite enough about why I believe its not problematic to be
> >different in this area. So why is the GNOME2 button ordering better?
> 
> The fact that it may be better is (unfortunately) not really relevant
> (and maybe if Apple wasn't so quick with their lawyers regarding look
> and feel this wouldn't even be an issue, which also brings up the fact
> that as Gnome2 appears to look extremely similar to the Mac will Apple
> also send its lawyers after Gnome?)
>
> When you are on a platform (Linux/Solaris) where everyone else does
> things one way you are asking for trouble when you decide to do things
> the opposite way. 

I don't feel you've adequately refuted my points on why it isn't a
problem to make this assertion. I still assert that there is virtually
no problem. Hence it *does* matter that there are substantial benefits
to the GNOME 2 button ordering.
 
> Though I guess it will make some people happy as it will make a
> Mono/.Net addition to Gnome unworkable given that any .Net apps will
> be following the MS style.

Using this argument we should follow MS style exactly for pretty much
everything, so we can make sure that .NET apps are also perfect GNOME
apps. Lovely, we're sure to progress that way.

> >The argument basically is that with left-right, top-bottom readers, your
> >eye is left resting on the lower right corner of a window when you are
> >done reading text. That makes the lower right corner the first thing you
> >read when you are done with a block of text. Thus it is the most quickly
> 
> Counterexample.  I briefly went into Gnome2 after reading this and
> ended up (without realizing it) taking a closer look at the UI.  When
> I logged out of Gnome2 up popped a dialog box asking me what I wanted
> to do, and it presented a list of 3 items (all short tiny little words
> with lots of empty space to the right).  Having finished reading the
> text my eyes ended up in the lower left simply because the right half
> of the dialog box was an empty waste land.  In other words, I had to
> go and search out for the OK button.

I fail to find your anecdotal evidence at all convincing. I could give
lots of contrary evidence myself, but I withold it because its very
subjective and easily biased by my already having taken a stance in this
discussion. Additionally, I already conceded that this point has been
open to a lot of criticism, and I don't believe our arguments are at all
premised on it.

-Seth