Re: [Usability] Re: Button ordering

[Kenny Graunke <kenny whitecape org>]

<quote who="Alan Cox">

> > "Show me the research" or "show me user testing" is a throwaway phrase
> > that is incredibly hard to argue against but isn't really helpful to the
> > situation. We don't *have* the ability to perform the research for most
> > usability choices, and in many many cases the necessary research is not
> > widely published because its done by *companies* with an interest in
> > being the primary ones to gain the benefit from the information.
> 
> So you are guessing ?
> 
> The point is that you are proposing a disruptive change. If the best
> argument you have for a disruptive change is that its what was recommended
> by two minor OS's while currently we generally follow what the mass of
> people expect then its not much of an argument.
> 
> If apple did the work, presumably the research is out there, otherwise
> perhaps Apple just guessed too. Or perhaps they picked one because it didnt
> seem to matter much.
> 
> Let me take a historical item still relevant today.
> 
> Everyone knows QWERY[Y|Z] keyboards are dumb. Everyone who tried to change
> the world and rewrite it  basically went out of business.

It's not like GNOME is going to go out of business. In fact, unless GNOME is
sufficiently /better/ than other systems, users will have no reason to use it
instead of those other systems.

Other systems have made good decisions and bad ones. If GNOME doesn't look
at these and use these experiences to aid its design, we're never going to
get anywhere. If you want something easy for Microsoft Windows users, clone
Microsoft Windows. Then people could switch - but would they? It might be
techincally better, but the software would still be pretty disgusting, and
it wouldn't have Microsoft's blessing. We can do better than this. Blindly
copying any system - whether Microsoft's, Apple's, or anyone else's - is just
plain stupid.

I don't want a monopoly to be controlling usability decisions for GNOME.
It's nonsensical. The GNOME Project is its own entity free to do what it
feels best for its users in the long run. Moving from Microsoft products to
GNOME is going to be tricky anyway - this is the *least* of their concerns.
There are many other issues which cause them far more trouble. Talking to my
friends who are Windows users, a multi-user system confuses them. This isn't
something that's going to go away, but it shows that there are more
important non-trivial things which present difficulties to them. For people
transitioning from GNOME 1.x, I imagine they would be relieved to at least
having GNOME consistent within itself instead of having inconsistencies
everywhere. I definately would be.

> If you were arguing at the point Gnome was being started I'd be all for it
> based on the limited info you have, but to change something that major and
> to change it to differ from the major desktop OS product and from general
> defaults for older gnome/gtk needs a much much strong argument.

This basically boils down to: "We can't change any decisions made two years
ago when our project was first starting - everything we've ever done is set
in stone." Seems silly to me - we certainly were not infallible back then,
especially in terms of user interface design. Just like any other issue,
things are worth revisiting when they can provide improvements.

We are *not* changing it on the basis of being like the Macintosh or unlike
MS windows. We're changing it because we in the usability project feel it is
better, not because we are a bunch of people who want to blindly follow the
Macintosh wherever it goes. We certainly aren't like that at all. Some of us
don't even like the Macintosh. Telling us that this is our only reason is
simply /dead wrong/.

Some reasoning for the layout:

When the user sees a dialog, they'll usually see the upper left and lower
right hand corners in a quick scan, before they begin to read it. Hence, the
most common action is in the lower right hand corner - where it will be
quickly noticed. If the user sees "Save" (a verb), and just tried to close a
document, they may just hit that before even bothering to look at the rest
of the "save before quitting" dialog.

This may be backwards for right-to-left locales. I don't even pretend to be
an expert on that, but I would imagine it is. If so, it should probably be
reversed via pango like other widgets.

This is a principle used in web page layout, marketing, and print media. Not
just user interface design. Furthermore, we have had people who are very
knowledgable about user interface design and observe many non-computer savvy
people in their daily lives. One of them works at an internet cafe, where he
observes people and sees all kinds of things most of us would never have
even thought of.

The other thing in our proposal is that it specifies that we should use
verbs instead of "OK", and so on - not just layout. This also will assist
the user. In the previous save example, the user sees "Save" right away with
this layout, so they know what's going on immediately, without having to
read all the text. "OK" could be something else.

> Which side of the road to drive on isnt a big issue. CHanging sides
> afterwards is...

Rather overdramatized analogy, I think. Lives aren't at stake here. I've
already adjusted to this dialog button order. It really does not take that
long. It takes me -far- longer to deal with the crazy button orders in GNOME
now - it is extremely inconsistent. Specifying this button order in UI
Guidelines and implementing it in GTK 2 will help make this issue go away.

The releases of GTK and GNOME 2.0 mark a time of change. If we're going to
change it, the sooner the better. We already have a patch, the usability
project has essentially come to a consensus.

> Alan

Kenny