Re: My comments on the WM spec

[Sasha_Vasko osca state mo us]

>---------------
>Owen wrote:
>
>> _NET_VIRTUAL_ROOTS
>>
>> Is this a) an (unordered) list of all virtual root windows other
>> than the real root window, or b) a list of the window that
>> corresponds to each desktop, either real or virtual?
>>
>> I'm a little nervous about including this since it is known
>> that using virtual root windows introduces a number of problems
>> for various operations (the ICCCM claims that a protocol
>> extension would be needed to properly implement virtual root
>> windows.)
>
>The ICCCM is right that this simple-minded approach to virtual roots
>can't work properly. If you want to know why, read Keith Packards
>work about a pseudo-root extension. It can be found at
> http://www.xfree86.org/~keithp/talks/proot.tar.gz.

The intentions of Keith are far more fetching then the idea behind 
_NET_VIRTUAL_ROOT. He talks about TRUE virtual root, with separate
window manager running in each, with client windows maping self on 
to the virtual root instead of real root, etc. 

The current 
intentions behind _NET_VIRTUAL_ROOTS are much simplier - virtual roots
are used only as a means to implement virtual desktops WITHIN single
window manager. From the point of view of client window virtual root
is merely a HUGE frame window, not a real root. All the messages are
still refering to real root window. Clients map themselves to the 
real root. There are no multiple window managers to install conflicting
colormaps, or cause any other conflicts. With enough care even clients
that use override_redirect windows (like KDE 1.0 for desktop icons)
works fine.

So all of what Keith Packard describes in his paper related to a 
completely different functionality, and has no relation whatsoever to 
_NET_VIRTUAL_ROOTS thing (except for name). 
Since he states that ICCCM is refering to 
the same set of functionality, we can assume that ICCCM statement that 
separate extension is needed, is not relevant for _NET_VIRTUAL_ROOTS either.

Yes, there are several drawbacks in using this approach, but those are 
much smaller, then what is involved in any other virtual desktop approach,
and are completely resolved by introducing _NET_VIRTUAL_ROOTS property.

>
>Matthias
>
>Sorry, I'm not on the list, please cc me if you reply.
>

Cheers
Sasha Vasko