On Wed, Nov 22, 2006 at 05:04:11PM +0100, Sander van Dijk wrote: > On 11/22/06, Anselm R. Garbe <[EMAIL PROTECTED]> wrote: > >Actually my favorit 'solution' to this problem would be, > >dropping the weights at all and don't touching the global client > >order. This seems to be more predictable (it also was the case > >until 0.9 afair). Though there are corner-cases as well, because > >zooming a multi-tagged client will zoom it in all tags it is > >tagged - when viewed. But to me this seems more natural after > >all. > > > >Actually I pushed this thing, it removes 40 lines, check hg tip > >if you like it ;) > > I think that's indeed more predictable. Also, let's not forget that > tags != workspaces. Caching views wouldn't be the way to go I believe.
Whether a wm is tiling or not is orthogonal to the tagging/workspaces distinction. So what's the important difference workspaces and tags? For me it's the idea that things can belong to multiple tags simultaneously, and in particular with workspaces (which typically only offer the choice of "on one workspace"/"stick on all workspaces"), I tended to be reluctant to move stuff from its original workspace temporarily because of the difficulty in moving it back to the right workspace in the right position. Partly because there tends to be a reason for the order so if it's not in that order you often end up rearranging things to be in that order, and partly because I'm still wondering if there's anything advantageous in terms of human spatial short-term memory in doing this. So, _if it is easily acheivable_, having different tag views being consistent with how they were last viewed would be good. Having said that, I spent some time last night thinking about maintaining different linearisations of a partial order and if there were any independences that could be used to simplify things and didn't come up with anything. Anyway, as someone with 25 tags I agree straightforward 2^n caching would be... interesting... from a performance perspective :-) cheers, dave tweed Send instant messages to your online friends http://uk.messenger.yahoo.com
