Newsgroups: comp.windows.x
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!noc.near.net!uunet!wyvern!taylor!mark
From: mark@taylor.uucp (Mark A. Davis)
Subject: Re: Blinking Cursor in Xterm???
Organization: Lake Taylor Hospital Computer Services
Date: Wed, 19 May 1993 12:45:43 GMT
Message-ID: <1993May19.124543.1948@taylor.uucp>
References: <1993May16.213936.6904@wariat.org> <me.737715054@dude> <1993May18.130845.6859@taylor.uucp> <1tbpi9INNfp3@early-bird.think.com>
Lines: 25

barmar@think.com (Barry Margolin) writes:

>In article <1993May18.130845.6859@taylor.uucp> mark@taylor.uucp (Mark A. Davis) writes:
>>You are stuck in a distributed system feedback loop!  What if you are on an
>>Xterminal or running xterm over the net to another machine?  That is when the
>>load problems occur.  If you had a machine with 20 Xterminals attached and
>>they all had blinking cursors in xterm's, this could represent a considerable
>>ethernet bandwidth waste.

>Ethernet certainly has enough bandwidth for a few blinking cursors.  My
>Lisp Machine blinks the cursor twice a second and updates the time in the
>status line every second.  This uses under 1K bytes/second, or about .1% of
>the bandwidth of Ethernet.  A hundred of them on a single ethernet might be
>a problem, but a couple dozen should be fine.

Granted it's nothing to loose sleep over, but this is Ethernet's tragic flaw:
the more activity (especially lots of tiny activity), the more collisions
happen and the performance gets exponentially worse...  I am just now
opposing ANY kind of waste of bandwidth under Ethernet.  Although in a polling
system it would not be so bad.
-- 
  /--------------------------------------------------------------------------\
  | Mark A. Davis    | Lake Taylor Hospital | Norfolk, VA (804)-461-5001x431 |
  | Sys.Administrator|  Computer Services   | mark@taylor / mark@taylor.UUCP |
  \--------------------------------------------------------------------------/
