janssen@parc.xerox.com (Bill Janssen) (12/20/90)
Dave makes a good point: 1) Virtual roots remove complexity from the operation of moving groups of windows. Of course, this could be done in the client library as well. 2) Virtual roots allow the server to optimize batches of moves/unmaps/etc. Perhaps we could also do this by defining the semantics of a window group more precisely... 3) Virtual roots allow interesting manipulations of stock applications which do not understand more complicated multi-window paradigms. 4) Virtual roots are a very interesting technology for doing shared-X kinds of work. Bill -- Bill Janssen janssen@parc.xerox.com (415) 494-4763 Xerox Palo Alto Research Center 3333 Coyote Hill Road, Palo Alto, California 94304
toml@ninja.Solbourne.COM (Tom LaStrange) (12/25/90)
In article <JANSSEN.90Dec19191544@holmes.parc.xerox.com> janssen@parc.xerox.com (Bill Janssen) writes:
Dave makes a good point:
1) Virtual roots remove complexity from the operation of moving groups
of windows. Of course, this could be done in the client library as
well.
2) Virtual roots allow the server to optimize batches of
moves/unmaps/etc. Perhaps we could also do this by defining the
semantics of a window group more precisely...
Yes, Yes, Yes! In fact, the early versions of swm had to do no work at
all when panning the desktop. I did not even send a synthetic ConfigureNotify
event to each client when the desktop panned because the client windows
did not move with respect to their "root" window (my interpretation of the
ICCCM). But alas, I caved in to public pressure and now send synthetic
ConfigureNotify events to all clients so that clients built with
toolkits that assume the "root" window means the "actual, X server,
root" window continue to work properly.
I think multiple, moveable, resizeable, iconifiable (is that a word?)
virtual roots would be very interesting....
--
Tom LaStrange
Solbourne Computer Inc. ARPA: toml@Solbourne.COM
1900 Pike Rd. UUCP: ...!{boulder,sun}!stan!toml
Longmont, CO 80501