Re: [wmii] suggestions for the sanitizing of swapping

From: Uriel <>
Date: Sat, 6 May 2006 14:10:16 +0200

You guys are wasting your time, larswm already got swapping and
cycling(which is even more important) right many years ago, but Im not
even sure garbeam ever learned to use cycling in larswm, no wonder it
has never worked in wmi(i), but it was one of the greatest larswm

Just copy larswm in what it did right and stop reinventing square wheels.


P.S.: And no, I wont explain how larswm worked, just go read the docs
and try it yourself it is well worth it.

On 5/6/06, Sander van Dijk <> wrote:
> Hi Anselm, hi list,
> Since the decision has fallen in favor of swapping, I have been
> thinking about how to sanitize it, since the way it currently is is
> just not release-ready. The main problem with (keyboard based)
> swapping right now is its lack of predictability. Another "issue"
> related to swapping is the suboptimal behaviour of moving. Below are
> some suggestions that I think will lead to behaviour that is both
> useable, predictable and consistent:
> - Invert the taglabel of the selected client in "default" columns.
> With "max" and "stack" columns, it is always clear what client is
> going to be swapped with, with "default" columns not, because one
> can't tell which client is the selected one. The suggested visual
> feedback would eliminate this unpredictability.
> - Make horizontal (keyboard based) moving attach the client just below
> (or above, whichever is the better default) the selected client of the
> target column. The would make swap and move behaviour more similar
> (which means, less stuff to remember), since they both "target" the
> same point in the target column. Inserting below/above the selected
> client is a better default (well at least it is not worse) than
> appending at the bottom of the column, I think, especially since this
> is more consistent with swap behaviour.
> - Make mousemove insert above the client under the pointer if it is in
> the top half of it, and below if it is in the bottom half (this has
> been suggested by both me and Chris before, I'm just relisting it here
> for convenience :-). This would make mousebased move and swap more
> similar, again since they both "target" the same point in the target
> column.
> - Keep the bindings for swap up/down on both M-C-{j,k} and M-S-{j,k}.
> One could interpret this as inconsistency, but I think the gain in
> useability is worth it in this case (and yes I know I stated the
> opposite two days ago, hey people change their mind some times :-).
> I think these changes would lead to an improvement is useability,
> predictability and consistency, and we get to keep swapping. Though
> I'm personally still not convinced we need it, I agree that it's
> unnecesary to kick people in the teeth with it's removal so shortly
> before release.
> Greetins, Sander.
> Btw, maybe horizontal swapping (M-C-{h,l}) and horizontal selection
> (M-{h,l}) shouldn't wrap around the edges of the screen, since
> horizontal moving (M-S-{h,l}) doesn't do that either (hence for
> consistency is might be better to do no horizontal wrapping around at
> all).
> _______________________________________________
> mailing list
Received on Sat May 06 2006 - 14:10:18 UTC

This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 16:04:23 UTC