Re: applied patch

[Sasha_Vasko osca state mo us]

> ---------------
>
> On Saturday 18 November 2000 11:55, Julian Adams wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> >
> > On Sat, 18 Nov 2000, Cristian Tibirna wrote:
> > >On Friday 17 November 2000 18:01, Sasha_Vasko osca state mo us wrote:
> > >> > I would still like to see StaysOnTop added to the spec.
> > >>
> > >> Oh, please, No! please, not again!
> > >> We've agreed that StaysOnTop is evil, let's not start it over again.
> > >
> > >1) I don't remember to have had agreement on this. We had discussion,
> > > yes. 2) I fail to see how this is more evil than other somehow
similar
> > > things that are already in the spec (the sticky things)
> >
> > You're right - there was simply no agreement. Lets just not bring it
> > around again. _please_ It just leads to other stuff being re-discussed,
> > and at this point (IMHO) all that is for V2.0 (V1.x ?)
>
> Why have things like Skip(Taskbar|Pager), but not StaysOnTop?  The
argument
> against StaysOnTop has always been that it doesn't belong client side,
and
> should be done through window manager configuration.  This is quite
> frustrating, I agree, but why include thinks like sticky and
> on-all-desktops if we don't include StaysOnTop (the next logical step).
>
> I'm bringing it up *again* because this is something that shouldn't be
left
> out.
>


Here is the first message in previous disscussion on the matter.

http://mail.gnome.org/archives/wm-spec-list/2000-June/msg00011.html

 Those who are interested - please reread entire thread.

In short here are the summary of the problems with StaysOnTop,
that distinguish it from other hints :

1) Where there is a clear definition for other hints like Shaded and
Maximized,
there are no such definition for StaysOnTop :
     - If there are 2 windows marked as StaysOnTop - which one overlaps
which ?
     - If StaysOnTop window is moved onto another desk, will it overlap
StaysOnTop
        windows already on that desk ?
     - What happens to Sticky StayOnTop window when current desktop changes
and it moves
       onto new desktop without user actually focusing it ? Will it overlap
other windows ?
       or other StaysOnTop windows will overlap it?
     - If window can be overlapped by another StaysOnTop window - then it
really
        is not on Top ??? If all the windows are marked as StaysOnTop, then
the one,
       overlapped by others is really on the Bottom ?
2) Request for StaysOnTop is in fact imposing a policy on Window Manager,
requiring it
to place window at the specific position in the stacking order - the first
position. Thus it is violating ICCCM and general X Windows approach that
clients should not
impose any policy, but instead simply hint on what they want and humbly
hope, that this
is what they'll get. You cannot really offer this kind of a solution to the
user in regards
to StaysOnTop hint. If user requests StaysOnTop - he/she expects window to
popup on Top.
On there other hand you can't really rebrand it to be HopefullyStaysOnTop
as
this is nonsense.

3) Similar to (2) : If Window Manager indeed implements functionality of
StaysOnTop
it is likely to provide its own configuration options, so by allowing
client to play
with that as well, you create ambiguity, that is very confusing to the
user, and not
easy to resolve.

4) StaysOnTop is in fact part of much larger concept of window layering, so
by adding
this hint now, you are locking this specs from further development, which
is not a
good thing. No need to limit ourselves to the half-measures. Future
generations will
not thank us for that :)

5) There are standard X Windows means that allows you to request Window
Manager to
popup window on the top of the stack untill it is overlapped by any other
window, which
should be enough for most of the applications. The request is
XConfigureWindow of
course, with stack_mode = Above and sibling = None.


Now I've listed all of the above not to solicite a feedback, just to
reevaluate what's
been told in previous disscussion.

I also ask to postpone all the disscussions on that matter untill 1.0 specs
are
finalized, and work starts on 2.0 version of it. Hopefully I convinced you
that this is
the matter complicated enough, to rush into it.


> > Julian
> >
> > >--
> > >Cristian Tibirna     : ctibirna total net     :
www.total.net/~ctibirna
> > >PhD Student          : ctibirna gch ulaval ca :


Cheers
Sasha Vasko