[OT] How long before..

[Peter Alcibiades]

Ken Corey wrote
> 
> 
> ....If the HIG are not scientifically provable "usability", but simply 
> subjective statements, then how can we measure "usability"?
> 
> 

The enterprise is fundamentally mistaken.  We have to start by recognizing
there is no such thing.  One size does not fit all.  Different people,
different working situations, different configurations.  If you think about
some areas where usability standards do work, what they have in common is an
identical set of tasks and a requirement to be able to move from one machine
to another without disruption.  All cars now for instance have gear shifts,
gear locations and signalling controls in the same place and they all work
the same.  But we are all driving the same roads to the same rules.  A
halfway example from UI in computers is double click to open.  But, my
experience is that given a choice, at least a third of people will choose
single click in the file manager.  And yes, this does contravene HIG rules,
because it leads to some openings needing double clicks.  Contrary to the
myth, they adapt without any difficulty and positively prefer it this way. 
Not all.  But a substantial minority.

The right approach is not to try to define something called usability and
then try to implement it.  The right way is to enable the user to choose how
he wants to work.  

The OS News readership is obviously very unusual, but you can see this
diversity every time they run a feature on readers desktops.  The variety is
simply wild.  Yet all these desktops are the result of someone's seriously
finding them nicer and more usable.  Flux for instance is right for me.  I
agree with Richmond that XFCE is very nice - have that on a laptop.  But
Evolution has a large following and apparently the new Gnome has
enthusiasts.

The lesson is, there are preferences, and usability is not a useful concept
in the way its being employed now, because no preferences are better than
any others.  Its a bit like asking how we find the correct size or style for
shoes.  The answer is that the right size and style is the one that fits and
allows the wearer to  move as he or she wants in the environment they are
in, and a lady at a formal dinner won't want construction boots.

If you want to test an application, maybe the right test is user errors. 
Track them and see what's going wrong and change it.  In defiance of
professional advice, I once permitted users to edit a data file directly,
not giving them a more long winded but more controlled interface.  It was
faster and simpler but error prone. This was the source of lots of errors,
and it had to be changed.  Once I found that people were opening the same
application twice, and had to correct what was giving rise to that.  The
mistakes people make when trying to use what we have written is probably the
most valuable test.



--
View this message in context: http://runtime-revolution.278305.n4.nabble.com/OT-How-long-before-tp4653161p4653267.html
Sent from the Revolution - User mailing list archive at Nabble.com.