VNC server based on kdrive using damage extension?

[Mike MacCana]

On Tue, 27 Jan 2004, Jaymz Julian wrote:

> On Tue, Jan 27, 2004 at 12:58:15PM +1100, Mike MacCana wrote:
> > On Tue, 27 Jan 2004, Jaymz Julian wrote:
> >
> > > On Mon, Jan 26, 2004 at 11:02:51PM +0000, Mike MacCana wrote:
> > > > On Mon, 2004-01-26 at 14:17 -0500, Bruce Bauman wrote:
> > > > > We would like to build a lightweight X server which will operate
> > > > > "headless". The only display device will be a remote VNC device.
> > > >
> > > > Why use VNC for this? The X protocol is already capable of proving en
> > > > entire remote desktop on a distant PC.
> > > >
> > > > (One answer could be: you have lots of Windows clients, and unlike
> > > > Linux / Unix / MacOS, Windows doesn't come with an X display server).
> > >
> > > Because while X is good^Wvagely acceptable at remotey viewing a set of single
> > > windows over a medium bandwidth connection, it's horrifically bad dealing with
> > > entire desktops, especially over low-medium bandwidth connections?
> >
> > > admittedly some of the x protocol
> > > compression packages do make this somewhat possible,
> >
> > You've answered yourself here. It seems far too often people
> > take X as is without any protocol compression. There's some
> > brilliant stuff that's been done in the last few years, some of
> > which make X quite fast over thin pipes - particularly MLView DXPC.
>
> lets say I agreed with you that these were as good for this, which I don't,

There's quite a few others who've tried these tools and seem to think
they're up for the job (reports of Mozilla running comfortable over a 9600
connection are surprisingly common). VNC compression is simply not as good
as what X can currently do. But it is more popular among people who use
more systems that don't have X clients and servers.

Combined with a general lack of knowledge about Xnest and modern X
compression techniques, the rather inelegant solution of taking an
application specifically written to be  shown on the screen with one
protocol, capturing it again, and displaying it with another has become
popular. That doesn't make it right.

> but lets say I did, it would still be the wrong tool for the job - X is the
> correct tool for exporting a bunch of windows, but it's not the right tool for
> exporting a desktop to a dumb and/or unknown bitmap device,

So you're implying that a 'desktop' (ie, root window) is somehow different
from another window? Or that X isn't the right tool for exporting to a
'dumb' device (I'm pretty sure there's thousands of X terminal
installations that match this description).

Mike