Task-centric UI

[Thomas Pfeiffer]

On 08.10.2012 17:05, Marco Martin wrote:
> On Monday 08 October 2012, Thomas Pfeiffer wrote:
>
>> I've already thought about application integration as a future topic a bit.
>> What came to my mind was Björn Balasz' presentation at the KDE UX sprint
>> April 2011 about task-centered interfaces. The general idea current
>> application-centric desktop shells are not really compatible with users'
>> mental models. A user doesn't think "I want to start Kmail touch", but
>> instead "I want to write an email to Grandma". However currently she first
>> has to think "Okay, what application do I have to start to write an
>> email?".
>
> this is not a new idea at all, i think it was kindof tried in the past with
> various degrees of success, at least in the old old days with document
> oriented uis (kindof a subset of task oriented, not enough anymore) as back 
in
> the days as apple lisa, early next, beos.

Yes, describing document-oriented UIs as a subset of taks-oriented ones pretty 
much nails it. A really task-oriented UI goes a bit beyond that.
But sure, the idea is not new, since it makes sense from a user's perspective.

> one thing that always killed it is a "branding" issue, is very difficult to
> make this work well with 3rd party developers.

Of course it killed it in the commercial world. Commercial developers want 
_their_ application to stand out, regardless of what's best for the user.

> The iphone-esque "apps" approach presents really the path of minimum
> resistance for 3rd parties, little difficulties and a lot of freedom for
> developers (weird to hear about freedom on the iphone, but being an app a
> little universe on its own, in this case applies)
> But is the best way for users? not so sure

Exactly. It's ideal for commercial 3rd party developers but fragments the user 
experience.

> we should learn from the past mistakes, and *maybe* we can pull it off, since
> we are in a considerably legacy free environment. We are also free to try 
this
> because since we really can't win the "Apps" game, we can try to just play a
> different game ;)

That's precisely why I think we as a thriving Free Software community are in a 
far better position to pull it off than the players in the 
proprietary/commercial world.
I don't say that we should force every application to fit into our task-centric 
scheme, but even with applications from the KDE community only we could 
already come a long way toward a seamless task-centric experience for most or 
all common tasks. If 3rd-party-developers want their application to stand out 
instead of fitting in, they can still make it so. But they should not come 
whining if users prefer the streamlined experience of community applications.

> (i would love the desktop going in this direction as well, but this is
> another, long, complex story :p)

I'd love that too, but that would be the long-term goal. I think Plasma Active 
is an ideal testbed for new ideas (see SLC for example) which can then move 
over to the desktop if they are successful.