Path: utzoo!attcan!uunet!lll-winken!sun-barr!newstop!jaytee!helium!db
From: db@helium.East.Sun.COM (David Brownell)
Newsgroups: comp.windows.x
Subject: Re: A tirade about inefficient software & systems
Message-ID: <3215@jaytee.East.Sun.COM>
Date: 5 Nov 90 23:15:05 GMT
References: <9011051527.AA03775@armory>
Sender: news@East.Sun.COM
Reply-To: db@east.sun.com (David Brownell)
Organization: Sun Microsystems, Billerica MA
Lines: 22

> As for X versus other windowing environments, the X distributed
> approach is a good one so long as network bandwidth is high.  Other
> techniques, such as those used in NeWS, are better when bandwidth must
> be preserved, such as over telephone lines.

There's another area where alternate approaches (e.g. NeWS) can be better
than X11:  where LATENCY is an issue.  NeWS lets you address bandwidth
scarcity by changing the message encoding used for a session; it lets you
address latency problems by eliminating some messages entirely.

With NeWS, you can (for example) make the server handle rubber banding or
menu selection without any roundtrip messages, and hence without extra
network/os latency penalties -- latency is eliminated.  Can't be done with
the generic X11 protocol, though one might define extensions.

Potentially, this approach can let one application process talk to more
displays at once.  On the other hand, UNIX doesn't have many applications
today where that's an issue, and I agree that it'd be good to see some hard
data.  Is anyone gathering any?

- Dave
One of the Million monkeys ... see, here's my keyboard!