On Sat, 15 Jun 2002 11:49:27 -0700, Andrew Gillham <gillham@vaultron.com>
wrote:

> On Sat, Jun 15, 2002 at 12:46:53PM -0400, James K. Lowden wrote:
> > 
> > What?  For umptiump years we've had to listen while every X-poor
> > vendor explained why shared logic, remote display systems were
> > backward-looking dinosaurs, and now that Microsoft's second try at a
> > closed method to partly imitate that feature has been incorporated
> > into their base product, now people are wondering if X-based systems
> > can do the same thing? 
> 
> No offense, but I have to disagree with you.  Saying that X is
> equivalent to Windows Terminal Server is just plain wrong.  The correct
> comparison would be with VNC.  The biggest (IMHO) advantage of WTS is
> that if you get dumped on your dialup or some other event that causes
> your session to get disruped, you can reconnect _exactly_ where you left
> off.  E.g. the application state exists only on the WTS box.  This
> precisely how VNC works on Unix.
> 
> Suggesting that the X Window System's remote display capability is
> equivalent is not appropriate, IMHO.  If your X Window Server loses
> contact with the machine(s) running the X Window Clients, you're totally
> hosed.

I think it depends on how you look at it, really.  In some ways, it's
true, VNC is more like WTS than X is.  There's no equivalent to
pkgsrc/misc/screen for X that I know of. 

In my defense, Andrew, I never said, "X is equivalent to Windows Terminal
Server".  I said WTS is Microsoft's implementation of a remote [graphical]
display system.  I would say that all three are implementations of that
notion.  It makes me wonder what others there may have been (whole
systems, that is, not VNC predecessors like PC-Anywhere and the like).  

One last point, while we're OT.  :)  I think if we looked at the protocol
layer, we'd find great difference among the three.  X is quite low-level. 
VNC is monitoring the raster display and sending deltas over the wire. 
WTS does something a little different, more like X (I would say), but at a
little higher level, something like you might get if you wanted to have a
"remote Gnome" protocol.  

I think it would be interesting to measure the number and size of
transactions on the wire; that might be a more scientific way to decide
which one is not like the others.  

Regards, 

--jkl