Re: icewm hints, etc...

[Toshio Kuratomi <badger prtr-13 ucsc edu>]

On Fri, 15 May, 1998 at 04:23:08PM +0200, Jeroen van der Most set free these words:
> On Mon, May 11, 1998 at 11:15:52PM +0200, Marko Macek wrote:
> > I attached the current icewm hints plans. Comments are very welcome. I am
> > fully willing to make necessary changes and would really like that people
> > would agree on one set of hints. If somebody wants to add a clean extension
> > to current icewm hints to suit other gui styles better I certanly add
> > and suport them (icewm is strongly oriented towards CUA gui style. I haven't
> > ever used NeX?t or a Mac).
> > 
> [SNIP]
> > I welcome comments on icewm hints (WinMgr.h attached). The proposal seems
> > a bit large and involved, but I have tried to be thorough.
> > 
> In the header file you've attached I found a property that contained the
> number of the current workspace the application is on and a sticky bit
> property. This only allows for an fvwm-like one or all workspaces behaviour.
> In other words, a window is on exactly one workspace (sticky bit off) or it
> is on all workspaces (stick bit on).
> 
> For some window managers this information is not enough. At the university I
> use the window manager that comes with Solaris' CDE. This window manager
> allows you to select on which workspaces the window is. For example, I can put
> my netscape window on workspaces One, Three and Four. I can put an xterm on
> workspaces One and Two and an xclock on all my workspaces.
> 
> This is more flexible than the scheme you propose. Instead of a single
> workspace property and a sticky bit, a window should have a list of workspaces
> it is on.
> 
I think you're going to need both sticky flag and workspace tips.  And perhaps
the scheme will get very complex... let's take a look::

Window Managers can implement two types of Multi-Screen capabilities.  We'll
call them Virtual Desktops and Disjoint Workspaces.  Both of these are
available in fvwm2 if anyone cares to look.  A Desktop is metaphorically an
extension of the on-screen Viewport.  The user can place windows that are
partially on one screen of the Desktop and partially on another.  The user can
also change their viewport so they are looking at any part of the desktop, not
just a full grid screen as defined by the pager.
Workspace is a separate Desktop entirely.  The user cannot place an app half
off of one Workspace and expect it to show up on another; this would be
metaphorically like placing your coffee mug half off your desk at work and
expecting the other half to show up on the table in your meeting room.

In fvwm2, setting a window sticky makes the application "stick to the glass"
and thus it travels between all screens, desktop or Workspace.  In ctwm (and
dtwm), there are only Workspaces, and each window is allowed to select which
Workspaces it wants to belong to.  Although no WMs implement "Stick to the
glass in Workspace 1 & 3 but do not appear in Workspace 2 and do not stick to
the glass in Workspace 4" it is possible that one may someday (and it is
possible to do using convenince functions now.)  Our hints should allow any of
these to occur.

We need a hint for which Workspaces the app appears on along with a hint for
whether the app is sticky on that workspace.  WMs that don't use all of the
functinality of the hint can still utilize it to get their desired behaviour
(fvwm sticky= set all workspaces+all sticky. cde 1 &3= set [1 & 3]no sticky)

-Toshio
-- 
badger  \"The Difference between today and yesterday is not so much what has
@prtr-13 \ changed between then and now as what I hope to change by tomorrow."
.ucsc.edu  \~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~=~