Bazaar Explorer prototype showing suggested Bazaar menu for IDEs
Stephen J. Turnbull
stephen at xemacs.org
Sat Jun 6 09:44:40 BST 2009
Ben Finney writes:
> Russel Winder <russel.winder at concertant.com> writes:
> > If an application is going to get used by a large number of people as
> > a core tool, it has to be usable -- I mean usable here in the
> > human--computer interaction (HCI) sense, i.e. when usability studies
> > are completed, the application gets a good result.
>
> No argument on that.
I have to disagree on that; just look at the prevalence of Word and
Excel. These applications are *not* usable, not for any of the
functions they provide over and above Notepad. git also got
widespread takeup before it got GUI, becoming even more popular than
the enormously popular pastime of bashing git's UI. In both cases, I
think the critical thing is network effects: the forms I fill out for
travel and purchasing are Word and Excel docs, and if you want to hang
out with Linus and Keith, you gotta use git.
Certainly, a good GUI will help Bazaar to get casual takeup, you know,
the kind of usage that RCS does quite well. But in the medium term,
it needs to demonstrate advantages that lead people who have
decision-making power for largish networks of cooperating users to
pick it.
In other words, I'm not suggesting any change in the immediate
strategy. But it's better to have a mediocre GUI that allows access
to all of the most powerful features of bzr, than a slick one that
attracts casual users but forces power users to the command line on a
regular basis. And once the basic features of the GUI are in place, I
think tasks like documenting and improving looms, adding a quilt-like
feature, etc., are more important tasks for the core developers to
concentrate on.
More information about the bazaar
mailing list