[3.0] suggestion: inboard scroll bars

Lex Spoon lex at cc.gatech.edu
Sun Feb 11 23:22:28 UTC 2001


"C. Keith Ray" <ckeithray at home.com> wrote:
> [...]
> > Overall, I don't think we should cater too strongly to people who insist
> > on things looking familiar.
> 
> This is the kind of attitude that Cooper describes in "The Inmates Are
> Running the Asylum"

Well, I often think that *HCI* inmates have too much control.  :)

Yes, I was mostly joking.  The main point is, we ought to try very hard
to pick something that most anyone can use and then set up the initial
image with those choices.  Preferences are a copout.  If typical users
need to tweak preferences right after loading Squeak for the first time,
then something is wrong.  The initial setting should all be fine ones
already.  Indeed, maybe this is just the case with the scrollbar
settings; I don't know.


The rest of this message is on the secondary point: we should expect a
*little* bit of mental effort from new users.  If they are going to turn
their nose up because the scroll bars are in the "wrong" place, then
they are going to be turning their nose up about something else, anyway.
 Let's focus first on people who are willing to work *with* us.




> I'm sure Don Norman would say that flap-out scrollbars, when not visible,
> have "no affordances". I certainly hate the way they flap out on top of
> other panes in browsers, making resizing the panes almost impossible.
> 

It's true: they don't have good affordances.  Neither do
context-sensitive menus.  Yet I'd claim these are both good computer
interfaces, even so.  Why?

It seems to me that affordances aren't nearly as important for a program
like Squeak as they are for things like toasters and doorknobs. 
Affordances are important for tools which you use infrequently,
especially if you aren't going to read a manual (or a Play With Me). 
Furthermore, affordances are easier to supply when the functionality is 
simple.  Squeak doesn't match any of these: people will be using it a
lot, they will spend time studying it, and the functionality is very
complicated.  For Squeak, it's okay to trade off affordances for
functionality or elegance.

I like the pianos vs. stereos analogy in this article:

	http://wex.www.media.mit.edu/groups/el/Papers/mres/pianos/pianos.html

Squeak is like a piano.  It's okay a piano's sustain pedal takel 30
seconds to learn to use, because it's going to take 30 years to learn to
use it well.  It's much more important that the functionality be
designed well.  The concerns are different than those for a volume knob,
where a 30 second learning curve would be ridiculous.

So, instead of affordances, let's talk about functionality.  Your point
about resizing being difficult when you use popouts, is a strong
technical argument that has nothing to do with how wierd many people
think they look.  I wish the discussion were more about things like
*that*, and less on how obvious or intuitive things will be.  I really
don't think we should worry if someone has to read a paragraph in order
to use the UI.  (I'll worry more if it takes them an hour, but I'd be
really surprised.)


-Lex

PS -- the best way to scroll is probably to use a special gesture with
your stylus  :)





More information about the Squeak-dev mailing list