User friendly Free Software Desktops

[home at alexhudson.com]

On Wed, Jul 18, 2001 at 11:24:43PM +0100, MJ Ray wrote:
> > the Debian installer has had a problem on every machine I've installed at
> > work, and seems to still have some bugs to work out (no, I haven't reported
> > them, yes, I know I should, I have them written down somewhere.. :( )
> 
> Yes, you should.

I know, I know :) In my defence, I am specifically talking about a r3
release I picked up at Expo a couple of weeks ago - I'm not sure how kosher
a r3 it is (although it is Cheepbytes), so I need to check that out too. In
Debian's defence, it is the biggest problem I've had with it so far - every
other release I've used (2.2 and before) hasn't had these issues.

I'm not sure what kind of state the Debian installer is in either - I was
under the impression that the current one was basically mothballed in
waiting for a new rewrite, but then you see them having problems with
bootdisks and wonder...

> Is this natural or symptomatic of the way technology has been
> introduced so far?

Cooper's argument is that IT is qualitively different from any industry up
to now, and that traditional engineering as we know it is actually failing.
In terms of 'real' engineering, it's a bit like the difference between being
a car driver and a car mechanic - but in the IT world, the cost difference
(i.e., constraints preventing you being a mechanic, even though you're
interested in cars) is much smaller than traditional industries, so the
polarisation between the two types is much more obvious. 

> I think there is space for a third sort of user, probably more of your HS
> than HL, who is willing to understand *why* something works, just so long
> as it works to achieve their goal. That's just a castle in the clouds,
> though.

I think to an extent that describes the 'Homo Sapiens' user in certain
conditions. One term Cooper is particularly in favour of is 'the principle
of conmessurate effort' - that is, getting out what you put in. In fact,
this is how he defines 'easy to use software' - not software that is so
ridiculously basic and simple that there's very little to understand in how
it works (the reasoning being any fool could then use it :), but software
that the user gains knowledge from using. So, they may have never used a
particular function before, but perhaps they are able to use previous
knowledge of how the program works to begin using the function. As they go
forward from this basic starting point, the user may need to learn new
things, but so long as the learning curve is constant difficulty as the user
progresses in knowledge, the argument is that they will be willing to put in
effort and gain knowledge if they see visible returns. The reason this
doesn't happen to a large extent currently is therefore due to the large
learning curve - the user puts in effort, but doesn't see return, and gives
up.

> > I think Free Software developers (who are, by nature, mostly Homo
> > Logicus, I would guess) need to understand the nature of software use a
> > lot better.
> 
> Probably.  You've said what it isn't and given properties of software
> that has got it right.  So what is "it"?

"It" I think is a number of things. Software that is easy to configure and
install in the first place obviously helps. Software that is documented
properly, not just for the developers, but for the intended user, not only
documenting the functionality and configuration of the software, but also a
task-based or goal-oriented help system (which wizards also lend themselves
to readily). An interaction design which is easy for normal usage (that is,
the 90% of work which usually uses probably 10% of the software's features,
the 'common tasks', should be easily accessible and shouldn't slow the user
down as they become more adept). A common task paradigm - like in Word, if
you wish to apply a style to a block of text, you highlight it and then
choose the style you wish. Other operations on the block of text work in a
similar manner. By using the mouse, the user is drawing the software's
attention to certain things - by highlighting the text, the user is telling
the software they wish only to operate on that area of text, for example.
The user should be aware of how to communicate certain facts to the
software, and in return the software should not ignore that, no matter what
the circumstances. I could go on.. 

In terms of Free Software, I think we do have a special advantage, above
Microsofts huge spending on focus groups and design consultants, and that's
the competition between different software, usually with different ways of
doing things. Although we're actually pretty piss poor at designing
interaction currently as a community (which I think everyone would admit??),
once interaction becomes a competition issue (when people start using the
app widely, non-geek users that is) evolution will begin to take care of it. 

Cheers,

Alex.


--