more efficient than X? (was RE: 'GO-Global' Related Commercial Ef fort)

[Alex Nicolaou]

> Can't think why anyone would...  This is supposed to be MORE efficient
> that X-Windows.  If you can't run over 10Mbps, then we're in 
> deep trouble
> :)

Ok, I read this and decided I must be doing something wrong. I then sat
down and read all the HTML pages in the docs tar (when all else
fails...) and tried out the X-server version of VNC (I don't have perl
but found the server was trivial to start directly anyway). 

While it is true that the X version is great, certainly adequate
performance and quite usable on my PC viewer (as the documentation
suggests) it doesn't equal the performance of my PC's X-server. It is
perhaps twice as slow at listing /bin inside a shell (kind of hard to
guess exactly, the operation is too fast to time with my usual X-server
and only barely slow enough to try timing with VNC). Have I missed a
step that'll make it go faster?

As for the windows version, the documentation quite frankly admits that
it is too slow, which I'd missed in my initial excitement to get it up
and running. It isn't too clear to me how it is implemented since I
haven't cracked open the source, but the implications of the WWW page is
that the WinVNC server is busy watching all the messages and trying to
figure out what's going on based on this combination of event watching
and polling sections of the screen. Microsoft NetMeeting does an ok job
of sharing apps even across slow connections (I use it over my internet
connection all the time) -- is it possible to build a VNC server on the
same API as NetMeeting? (I don't mean using the netmeeting  SDK which
probably requires netmeeting on the other side; I mean using the same
underlying method that netmeeting uses to support remote application
control).

alex