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

List:       kde-core-devel
Subject:    Re: While were making controversial suggestions... ;)
From:       Michael Matz <matz () ifh ! de>
Date:       2000-04-28 23:15:00
[Download RAW message or body]

Hi,

On Fri, 28 Apr 2000, Kurt Granroth wrote:
> [... desc. of fitts law ...]
I fully understand that.

> As a quick example, the mac menubar does a terrible job of being close
> to the cursor BUT does a phenomenal job of being big.  In fact,
> because it is anchored on the top of the screen, it has INFINITE
> height.

Yes, virtual infinite because it not taking more space on screen as it
should (read below). But note, that one doesn't need this infinite height,
as the larger the normal menus are in height, the better they are also to
select. So there exists a height (for each user different) over which it
make no difference if I add another millimeter. Whereas for the distance:
the larger it gets the worse.

This can be clearly seen in the hypotetical case of an 100" screen, with
normal mouse and menu on top. You can say what you want, the height is
infinite, but neverthelesse the distance is far too much as that fitts law
would hold. I guess in my experiment I was exactly getting this break even
point as I have measured the same time for using the according to fitts
law better to target macmenubar and the theoretically worse normal
menubar.

So my conclusion is that fitts law only holds in some circumstances (of
course), and that menubars are not necessarily one of them, or at least
they are a border case.

> But back to the the B II titlebar.   B II does a terrible job of
> Fitt's Law in *both* areas.  In the example I gave, the cursor will
> necessarily be on the right of the window to give it focus while the
> titlebar will be on the left making sure that the cursor isn't
> anywhere near the titlebar.  Furthermore, unless you have a very large
> title, the bar will be quite small, making it harder to select.
> 
> A full toolbar fits both criteria by being larger and right under the
> cursor.

This is all true, and yes B2 title bars break fitts law, as you presented
it. But my point was here also that fitts law is not really
useful. For that to see, one should ask, when I want to apply fitts
law. I think for UI elements which are used often, and under the premise
that I have space. I count title bars as not often used, so for me fitts
law is not useful, and consequently I don't care if these title bars break
fitts law or not. You obviously do care, and about that we can do nothing.
(then again even with normal title bar, to make them better fit fitts law
we should make them at least 50 pixel high. Now this is clearly wrong, but
with only fitts law in hand we would have to do that.)

But only for another example why fitts law is not in every case usefull:
If we would apply it to each UI element very rigorous, we would have to
make them very big and close together, for that we would use an hexagon
pattern for the active UI elements as they have the best space partition
without any space between. Obviously that can't be the solution. For once
without space between we can accidently select the wrong thing, second we
might be resticted in space to share, so we can't make them as big as we
wish, and third, all items have the same size, which is plausible but may
be we have checkboxes and buttons, and they should have different size.

It can be seen, that to make fitts law usefull, one also has to take into
account space to share, importance of the item, astethic issues and so on
(not to speak about the wrong infinity above), not everything of it can be
measured and pressed into figures. Then we would have to apply this law to
_each_ and _every_ UI element, buttons, icons, checkboxes, whatever, and
globally minimize (or maximize if you will) this function.

I don't doubt that fitts law has its use, as long as it's not slavish
applied, but it's definetly not the medicine to heal every HIC issue. It
has so many flaws that one should think two or three times before doing
something with it. Sometimes normal human sense is enough to reach the
same, where fitts law would apply, and at the same time also does the
right thing where fitts law would break horrible.

All that said, I admit that B2 breaks fitts law, but I think the users can
cope with that, like they can cope with it in Be, which is really used by
'users' (mainly multimedia artists).

> > his life. We could even explain why users might want to choose this or
> > that (talking about pros and cons).
> 
> That's not an option.  If we FORCE the user to choose one, then the
> initial startup would just be a pain in the ass for the majority of
> users.  If we don't force them, then *something* needs to be the
> default.

Hmm, I think mosfet (or?) said a twm/athena look is in order. That could
be the default, as it is so ugly, that every user would choose something
else. So effectively we force them to choose something, while at the same
time nobody can object, as we really have a default ;-)

Just kidding.


Ciao,
Michael.

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