Re: Regressions in GNOME 2 vs 1.x

[jacob berkman <jacob ximian com>]

On Fri, 2002-02-08 at 14:08, Alex Larsson wrote:
> On 8 Feb 2002, jacob berkman wrote:
> 
> > On Fri, 2002-02-08 at 07:35, Trever L. Adams wrote:
> > 
> > > Second, the task list (Again edge panel), groups any given number of
> > > windows from the same task together.  I liked the old behavior of giving
> > > a minimum before it will group.
> > 
> > i agree here.  when i use the task list i like it to group when there
> > are > 1 tasks in the group.
> 
> You mean you want it to always group?

yes.

> Currently the options are 
> a) never group
> or
> b) try to be clever
> 
> The algorithm for b is somewhat involved, and not quite finished.
> Basically, it tries to give each button at least DEFAULT_GROUPING_LIMIT 
> pixels, and if that is not possible it picks one group to group and tries 
> again. picking which group to group next is done by a scoring function 
> that gives a score for each group depending on factors such as:
> * Number of windows in the group
> * Number of windows in the group with the same title 
> * Total abount of focused time (this is not implemented fully yet)
> 
> The final scoring function has not been tuned at all, and probably should.
> 
> Perhaps we should add a third option
> c) Always group windows

i think this additional option is neither confusing nor bloated, and i
think the 3 provide a good balance of what users would want.

jacob
-- 
"In fact, can you imagine anything more terrifying than a zombie clown?"
	-- moby