Scoping the UI changes for 2.0

[Ian Clatworthy]

Martin Pool wrote:

> These are really interesting and important conversations but I don't
> want them to dominate getting things finished that are already almost
> there.

Agreed. In some selected cases (like branching outside a shared repo),
the performance problem is actually a UI bug though, so I feel that
is absolutely in scope for 2.0. We simple can't take 4 minutes to
create a local branch when 'proper' use of the tool does the same thing
in 4-5 seconds (tree per branch) or less than a second (shared tree).

> We want to get back to a clean and cohesive
> model, and I think that may come from looking more systematically
> across all of the cases we have to cover, and the constraints or goals
> on them from both the user and performance sides.  Before we settled a
> specific design for the new repo format we had a list of
> characteristics the design had to meet and that seemed to work well.

To capture the conversation to date and provide better structure going
forward, I've created an umbrella spec with an initial set of desirable
design characteristics. See http://bazaar-vcs.org/DraftSpecs/BetterNewUserUI.

With regard to the performance vs UI focus ...

My primary objective this weekend has been to get the initial release
of bzr-historycache out (announcement pending) which does nice things like
speeding up log of a single merge revision from 30 seconds to 1 second for
Emacs. I've also put some effort into these UI specs though to clear my brain
before taking a few days leave. :-)

Ian C.