More on GTK usability



Hmmm, I do seem to have started quite a thread here.  

It seems that there are a stack of relevant observations:

Someone made the excellent point that users expect consistent
behaviour across applications on their chosen platform.  Now for me, and
many other desktop *nix users, the platform model we have is
the shell.  Having the gtk1 chooser behave like a better version of the
shell was great.  

Of course, for the "desktop converts" that GNOME is trying
to win, this does not apply and will never apply unless someone sits
them down and and takes the time to show them how to do funky stuff with
the shell.  So there is an inherent problem: you don't really know what
the "platform" for GTK is.

If you're going to assume it's a serious lowest-common denominator, a la
Mac OS (which I've used a lot and find to be slow and awkward), then
please just make a way for us old school shell types to keep our
platform as an optional setting.  We'd even be happy to edit a config
file somewhere to turn it on :).

I think the theoretical best outcome would not be this "two personality"
solution but rather: something that is 100% compatible with shell-based
user behaviour, with all of its quirks (control-u to blank the line,
anyone?) but which simultaneously offers new users an intuitive way to
discover all of the same powerful functionality.

Or alternatively, I'm willing to sacrifice 100% compatibility and learn a
second set of conventions for the sake of converts if I have to,
provided that the new interface is as efficient as the shell (efficiency
= keystrokes per task).  GTK2 isn't.  I'd nominate the openoffice.org
solution as a good example of something that breaks a few shell
conventions, but does so consistently and without sacrificing much
efficiency (though it doesn't handle dotfiles perfectly, that could
easily be fixed).

If the *nix desktop is moving in that direction, someone should really
get around to building a new command line shell that is GUI-fied and
does things the same way.  But that's another project :)

-- 
Peter Eckersley
Department of Computer Science   &                  mailto:pde cs mu oz au 
IP Research Institute of Australia             http://www.cs.mu.oz.au/~pde
The University of Melbourne               



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]