the next step on the desktop

[todd rme]

On Thu, Feb 3, 2011 at 12:16 PM, Ivan Cukic <ivan.cukic at kde.org> wrote:
> A.S. I might sound like I'm totally against the ideas proposed, I'm not -
> I'm just pointing out the potential problems instead of just +1-ing the
> good parts.
>
>> That doesn't seem to be a problem with the add widgets dialog, or with
>> the default Dolphin icons view, either..  I think it is fine if they
>> are clearly separated visually.  These aren't big blocks of text, they
>> are clearly separated combinations if cons and text.
>
> Lets entertain the idea that it doesn't (although I'm not lenient to agree
> on that one :) ) - you suggested to show it in the panel itself. By
> default, panel's height is not enough to fit in the 'icon above text'
> items, you you'd need to end up with a format like the current taskbar.

Right, which is why I said "replaces or covers the panel", rather than
just "replaces".  If the panel is big enough, it would replace the
panel, but if it isn't you would need something bigger that would
cover it.

>> That assumes that you use the whole height of the screen for the
>> vertical menu, but do you?  Certainly not be default.  I think it
>
> Well, you can. If you use the classic menu, it stretches as much as it
> needs. Shelf applet as well (though it can't really be used for general
> app launching yet)

Right, but by default the default application launcher does not do
this.  This is meant as a default layout optimal for most basic users.
 A user who knows enough to resize the application launcher or switch
to a different one would also know enough to change out my proposal
for one of the existing menus.

>> That assumes you only use it for launching applications.  This would
>> not be the case.  It would be used for krunner, for instance.  I have
>> lots of trouble with krunner's calculator and dictionary runners since
>
> Ok, again, when you do the calculations, krunner takes 200x50 pixels
> (approx) - is it much? (I'm not saying that replacing panel contents is
> bad, just that it has some bad side-effects)

Those are only examples.  For the activities runner this will be an
even bigger issue.

>> If you are solely concerned with issue with lunching applications, you
>> could make it so that clicking one of the application folders brings
>> up a vertical menu that you can then use to navigate that folder.
>> Whether it displays a menu or replace the current horizontal view
>> could be up to the individual thing being displayed, since while I
>> agree some things are better displayed vertically, while others are
>> better displayed horizontally.
>
> The idea is intriguing. I have my doubts when reading about it, but it may
> prove to be perfect while using it. (if it becomes real)
>
> This is something that could go very nicely with the global menu bar (ala
> MacOS) - the horizontal menu you are proposing is essentially that, but
> for the 'workspace' and not for the app itself.

I guess somewhat.

>> This is a problem with any grouped task manager.  The default task
>> manager has the same problem when grouping is enabled (which it is by
>> default).  In an earlier email I explained this in the benefits and
>> drawbacks of dock-style task manager.
>
> Agreed (as with your pros and cons of docks). I just don't find the
> 'current default has problems' to be a reason for 'it is not important
> whether a new idea has the same problem)

My point was a comparison between the existing default and a new
proposed default.  This is not a drawback from this perspective since
it is the same.

I agree this is a drawback, but I think it is more compensated for by
the benefits I listed.  If there was a way to make it work in all
situation I would be all for it, but I have not seen and have not been
able to come up with a solution that has all the benefits of the
existing interfaces.

>> True, which is why I am proposing putting more focus on favorite
>> applications and krunner, putting both of those right in the panel.  I
>> think accessing those without having to bring up the application
>> launcher would be much faster.
>
> I agree, but that part is possible with the current system as well. Just
> for that, we'd only need a krunner plasmoid (IIRC, it exists, just don't
> know where :) )

We would need a krunner plasmoid, and a better interface for assigning
launchers to the task manager, and I think a better way of displaying
launchers, and I think it needs changes to the application launcher as
well (such as removing favorites, and putting more focus on finding
applications you don't use often).

I am not proposing just changes to the interface, the focus is on a
change to the work-flow, with changes to the interface chosen to
encourage and support this work-flow.

-Todd