RE: Iconic/Minimized/Shaded poorly defined (was _NET_WM_STATE_MINIMIZED proposal)

[Elijah P Newren <newren math utah edu>]

On Wed, 2004-06-16 at 05:39, Giles Atkinson wrote:
> Elijah,
> 
> What is the semantic difference between setting
> _NET_WM_STATE_MINIMIZED and
> setting the ICCCM property WM_STATE to IconicState?  If there is none,
> this does not look like a good idea.

My understanding: _NET_WM_STATE_MINIMIZED implies WM_STATE should be
IconicState but not vice-versa.  Namely, WM_STATE is set to IconicState
when _NET_WM_STATE_SHADED is set as well.  (At least, Metacity does
this; if it's not supposed to, then I couldn't find anything against it
in the spec and the spec needs clarification).

The EWMH spec seems to imply there's a difference between these two as
well.  Under the implementation note for virtual desktops, it says that
"Windows which are actually iconified or minimized should have the
_NET_WM_STATE_HIDDEN property set..."  Although that also seems to imply
that minimized isn't a subset of iconified, which seems odd to me.

Honestly, I don't know the difference between IconicState and
_NET_WM_STATE_HIDDEN, so I'm curious why _NET_WM_STATE_HIDDEN exists. 
There's some implied difference between these two in the EWMH spec,
which seems both vague and odd to me: "If the..._NET_WM_STATE_HIDDEN
hints [is] not present, and the Window Manager claims to support
_NET_WM_STATE_HIDDEN, then the window should be shown if it's in either
NormalState or IconicState....For Window Managers that do not support
_NET_WM_STATE_HIDDEN, the pager should not show windows in
IconicState."  This doesn't seem to spell out the semantic idea other
than giving the actions necessary for this one issue and I really don't
see the need for this course of action in this corner case.  I
personally think of them as equivalent, but would welcome any
clarification on the subject.

> Also, is there any semantic difference between "minimised" and
> "iconic"?

I couldn't find a good definition of iconic.  I'll take it to mean "a
smaller representation of the window is visible which cannot be fully
interacted with."  In which case, both shaded and minimized satisfy the
requirements, making minimized a subset of iconic.

> The former looks like poor terminology for X as the window size is
> unchanged.

True...are you suggesting a search and replace rewording to the EWMH
spec?  If so, what would you replace it with?  If you want to use
"iconified", then that term would need clarification (and a different
definition than I gave above) and we would have to explicitly make
shaded not be a subset of iconified (which, at least for Metacity, would
be a change from the current behavior, though one which may make sense.)


Thanks,
Elijah