[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-look
Subject:    RE: One-Click", "Two-Click
From:       "Markus Holzem" <markus () holzem ! de>
Date:       1999-07-09 6:32:06
[Download RAW message or body]

> From: rik@rikkus.demon.co.uk [mailto:rik@rikkus.demon.co.uk]
> 
> >> Single left click == select.
> >> Double left click == action.
> >> Single middle click == action.
> >> Single right click == context menu.

I'll pick something out from the bottom of the mail:

> Conclusion: The idea at the top of this mail was plucked from the back of my
> head at random, but thinking about it more it seems like it does actually
> please all. The only irritation will be to some current KDE users who expect
> the left mouse single click. All they have to do is switch to the middle or dou
> ble click. I think the future user base of KDE must be the priority. If KDE 2
> isn't perfect, I'll be ashamed. KDE 1 is pretty close.

Somehow I'm disappointed that I write hundreds of lines and no-one seems to
read them. What came to your mind (mail at 18:36) was written as an option
by me about ten hours earlier (8:55) with explanations why it is diffucult
to have a standardized feel of the mouse AND single click on icons.

I have volunteered to write the KDE user interface guidelines. In the chapter
about the mouse interface I would have to say:

	left button click on unselected object:
		selects the object, but sometimes it performs the default action
		for the object
	left button double click on unselected object:	
		selects the object and performs the default action for the object
	right button single click:
		selects the object and shows a context menu

Doesn't seem so horrible, but if I write down for keyboard shortcuts:

	CTRL-N: opens a new document, but sometimes formats your harddisk

you wouldn't think it really made sense, would you ;-)

> I dedicated the last year of my degree to HCI (Human Computer Interaction) and
> ended up disagreeing with a large amount of what was held up as the 'truth'.
> 
> HCI purists are a bunch of idiots. They either:
>
> Have some grand scheme in their head for a new interface that in their humble
> opinion should replace all that's gone before because it's revolutionary and
> 'better'.
 
All purists are idiots ;-) I'm mostly vegetarian, but I eat meat; I don't own
a car but hire one whenever I need it; etc. Go to the extremes and no-one
believes you... THAT is one ultimate truth.

> Or, they have grand opinions about current software. Take the GUI Hall of Shame
> page (linked to from Troll Tech). They have some very good points to make but
> they're a little arrogant.

That comes perhaps from having to do with developers ;-) It is sometimes
hard to break the developers view of things, because he knows the system
that well, that he cannot understand what troubles the user. It is the main
reason, why a lot of HCI professionals depend on their VCR equipment, not
to evaluate the stuff but to proof problems to the developers, CEOs, or
executives.

I've developed software since I had this nice optional afternoon course
in school, more that fifteen years ago. I try always to maintain both
views and to step not on to many feet, but the idea of a single click
for action without an option to select is honestly very difficult to
understand for the user.

How often do users fire up kfm? Or Explorer? Or any other file manager?
It is a subtask. You edit files. You edit documents. You edit graphics.
You edit music. And everywhere else the rule is: single click=selection
and double click=action.

Exceptions from this rule come (and should be allowed), when there is
really nothing to select and selection and action look that similar
that the border becomes very narrow - like on hypertext systems.

> For example, they say that Microsoft's (?) idea of losing the raised border of
> toolbar buttons is 'wrong'. Their contention is that the buttons no longer look
> like controls. This is perfectly true, but:
> 
> On first look at the toolbar, the user sees it's in the same place as the
> toolbar normally is - it looks like a toolbar - they move the mouse over - it
> acts like a toolbar. From that second, their brain adds 'toolbars may be flat'
> and they continue.
> 
> A flat toolbar is actually a brilliant idea. Think about it - with less clutter
> on the screen, your eyes don't have to do as much work. You focus more easily
> on what you're working on.

I agree: the user goes for the icon and less (or even no) extra space is needed
to visibly separate the tools. The option "large icons" satisfies the needs of
untrained mouse users. And the visual feedback (showing the border) ensures that
he really did hit the spot.

The problem with MS is, that they don't change their environment when a new
release of MS Windows comes up, but they change it with any new release of
their software:

	MS IE 4: new desktop
	MS Office 2000: new file open dialog
	...

And due to the development environment and the error prone DLL system (no-one
wants to rely on MS DLLs) you have applications with various looks and feels
in your working environment.

And all companies and developers have to imitate the new look and feel, because
it simply looks good. The user has to buy the new releases of software to get
again a consistent look and feel, and when the user nearls succeeded MS finds
a new fancy look, or a new fancy widget, or... MS applications are fairly
consistent in their look and feel, but you don't run only MS applications.

This is why all users and companies bear a grudge against Microsoft. Give those
people a consistent environment that they can rely on, and you get them on
your side. Perhaps they even start to use Qt instead of Motif ;-)

> We have seen both camps talking here. I'm not complaining. That's what this
> list is for. Just don't be arrogant and assume you're right. There are no right
> answers, only satisfactory solutions that may involve compromise.
> 
> So I summarise:
> 
> To stick to principles for no reason other than that they're written down is no
> way to work and certainly no way to enhance the design of KDE, which has only
> recently become a completely usable desktop environment. It's now that
> innovations need to be made, not later.
> 
> To make your software different to all other just because you have a better way
> to do something (single click _is_ better as it's easier) is no good either.
> You must conform to standards otherwise you're dead. I'm an X old-timer. Using
> a Mac is painful for me. For a Windows user using KDE for the first time must
> not provide a revolutionary interface that takes too long to get used to.

I really can't see why the single click is that revolutionary. Revolutionary is to

* create a sensible bi-handed environment
* provide functionality that the users crave for
* integrate speech recognition consistently from bottom to top

If I suggested pie menus, everyone would ask me whether I was crazy. For each
option you have only to move a little bit from the centre. But they become
extremely large (more that twice the width of the rectangular menu) and they
are not equally usable for users of trackballs or other non-mouse pointing
devices.

For those who don't know them: menu items are not listed one below the other,
but around the mouse pointer. If someone is interested:

	http://art.net/~hopkins/Don/piemenus/ddj/piemenus.html

> The advantages and disadvantages of the above mapping (as default):
> Windows users: Behaviour is almost identical to what they're used to.
> KDE users: Behaviour is again almost identical. After a few minutes they'll be
> used to the new system. They'll also thank us for allowing them to select icons
> without activating them.
> Motif users: Identical behaviour (with extension on the middle mouse). The
> mapping of the middle button shouldn't interfere with the usual paste function.
> Mac users: Obviously anyone who uses a Mac only has one finger. They'll have to
> get some prosthetics.

Not to forget any non KDE-applications on any UNIX workstation. THEY RUN
ALONG WITH KDE. KDE IS NOT A SMALL PLANAT ON ITS OWN. IT IS A **UNIX** DESKTOP.

> Hmm. There was supposed to be a conclusion earlier.
> Anyway, don't think about yourself or your education. Think about your users. I
> have run networks with 1100 users. You get to realise that you have to bow to
> them, no matter how crap they are at reading manuals or remembering mouse
> buttons. Add enhanced functionality for yourself but don't make it the default
> and if it interferes with the standard, make it a non-default configuration
> option.

Couldn't have said it better.

Ciao,

Markus

[prev in list] [next in list] [prev in thread] [next in thread]