Pleasant@Rutgers (09/01/82)
HUMAN-NETS Digest Tuesday, 31 Aug 1982 Volume 5 : Issue 90
Today's Topics:
Queries - What are Your Favorite TV Shows &
How Do I Mail to this USENET Site,
Programming - Games and Heuristics (2 msgs),
Technology - User Interfaces & Print Fonts (3 msgs),
Computers and People - Computer Names &
Motivating non-Technical People to use Computers
----------------------------------------------------------------------
Date: 31-Aug-82 14:20-PDT
From: JWAGNER at OFFICE
Subject: Hackers' favorite TV shows -- survey
I would like to compile a list of hackers'/programmers'/engineers'
favorite television shows to see how the list compares to ratings
by Nielsen and other services (Arbitron, etc.). Current programs,
reruns, cartoons, whatever your favorite is, I'd like to include it
in my survey.
If your job is in a related field, I'd like to hear from you.
Students taking CS, engineering or related courses (or their
instructors) are welcome to respond.
Please send along a very brief job description with your TV
favorite. Results will be made available when the survey is
complete. Please send all responses directly to me,
jwagner@office. Thanks.
Jim Wagner
------------------------------
Date: 31 August 1982 18:00-EDT
From: Robert Elton Maas <REM at MIT-MC>
Subject: how to get message thru???
We've been trying for weeks to get mail thru to:
menlo70!sytek!zehntel!teklabs!tekcrd!tekcad!keithl.at.UCB-C70
Does anybody know how to reach this mailbox on USENET?
------------------------------
Date: 30 August 1982 21:20-EDT
From: Phillip C. Reed <PCR at MIT-MC>
Subject: Distributed Games
There was an article in BYTE a while back referencing a game that
was played between two PET computers that are wired together. I
believe that it is called FLASH, and that it amounts to a tank
battle, where each player can only see the terrain near his tank (as
modified by tree lines, hills, etc.).
Granted, this isn't really a network...
...phil
------------------------------
Date: 31 August 1982 1931-EDT (Tuesday)
From: Craig.Everhart at CMU-10A
Subject: MazeWar game for Altos
There is no centralized service for MazeWars; in fact, I think it's
impossible to play the same game on different Ethernets, since the
packets it uses aren't Pups, and therefore aren't transmitted by Pup
gateways.
Remember, too, when you play with possible distributed architectures
for this and similar games, that Ethernets only deliver packets with
high reliability, not with perfect reliability, so it's usually
simpler to use some architecture other than token-passing (where
you'd have to take special precautions against the token getting
lost). I believe that MazeWars runs by each player's Alto sending a
packet for each move made, picking up on packets telling where all
the other players are, and having some "I'm shooting you" protocol
between battling players; the shootee has to agree to die. But each
game will time out the existence of other players if it hasn't heard
from them.
As far as adding net and host traffic and overhead, you pretty much
can't stop taking up the net (even though you're operating at
keystroke speed and sending only tiny messages), but you can play
tricks to keep hosts from having to discard unwanted broadcast
packets. For instance, you can set most 3Mb Ethernet interfaces to
receive packets to any one host, or to receive "promiscuously"
packets to any host (this is how you write protocol debuggers).
Usually you set your receiver to pick up only those packets that are
addressed to you; but Trek does a clever thing. For each universe
being run, it computes a set Ethernet address; there are only about
15 different universes it can run. So when you join universe N, all
copies of the game program in the various machines compute Ethernet
address K = f(N), and set the Ethernet receivers to receive only
packets addressed to K, and send all game traffic to (simulated)
host K. This produces a directed broadcast! And most other
receivers on the net will ignore these packets, because they're
directed to some other host. On the Unibus 3Mb interface, no memory
transfers or interrupts happen for such packets.
Key points: truly distributed control in light of only
mostly-reliable transmission (as well as people dropping into and
out of the game at any time), and this directed-broadcast trick.
Craig Everhart
------------------------------
Date: 30 Aug 1982 2207-PDT
From: Les Earnest <LES at SU-AI>
Subject: Recognition of cursive writing
Regarding the handwriting recognition scheme used in Applicon cad
systems that is based on a "paper from lincoln labs in the late
50's" (H-N V5 #87), I believe that the latter was one I wrote.
While the original paper is inaccessible, an accurate description
can be found in IEEE Spectrum, May 1965, "Machine Recognition of
Human Language, Part III -- Cursive Script Recognition" by Nilo
Lindgren.
I am gratified to learn that someone is making use of this work --
it was news to me. To my surprise, no one seems to have developed a
more reliable scheme in the intervening 20-odd years.
------------------------------
Date: 29 August 1982 10:08-EDT
From: Zigurd R. Mednieks <ZRM at MIT-MC>
Subject: HUMAN-NETS Digest V5 #89
I can readily believe that, depending on the quality of font
information available at a given site, one might prefer Helvetica to
Times Roman. For instance, the font information that comes with
Berkeley VAX Unix is awful, and this lack of quality might affect
serif fonts more than sans serif. Even though serif fonts were
designed to be readable after the degradation involved in printing,
they may not stand up as well to the degradation from poor
digitization.
There is yet another issue here: Times Roman was designed for narrow
newspaper columns. Using it in a paper laid out as a single column
of text would make that paper hard to read. I am not advocating the
abandonment of Times Roman -- in fact I would rather see more papers
use Times Roman and a two column per page format instead of a less
condensed roman font like Hershey and single column format.
Cheers,
Zig
------------------------------
Date: 29 Aug 1982 1137-PDT
From: Lynn Gold <FIGMO at KESTREL>
Subject: Helvetica vs. Times-Roman
In the end, unless you have to deal with copiers which don't like
certain fonts (I once worked on a paper where a font called Broadway
didn't show up very well), it's all subjective.
When I'm choosing a font, I tend to go by a combination of what is
available on my output device, what looks good/best out of my
possible choices, and what I'm printing.
--Lynn
------------------------------
Date: 29-Aug-82 19:44:22-PDT (Sun)
From: allegra!rba at UCBVAX
Subject: How printing affexts readability
There is an extensive body of research on how various
characteristics of printing affect readability. Two
references relevant recent human-nets discussions are:
A.J. Campbell, F.M. Marchetti, & D.J.K. Mewhort,
Reading speed and text production: A note on right-
justification techniques. Ergonomics, 1981, 24, 633-640.
P.A. Kolers, R.L. Duchicky, & D.C. Ferguson, Eye
movement measurement of readability of CRT displays.
Human Factors, 23, 1981, 517-523.
Bob Allen BTL-MH
------------------------------
Date: 30 Aug 82 13:18:33-EDT (Mon)
From: Andrew Scott Beals <andrew.umcp-cs@UDel-Relay>
Subject: network naming
I can think of 2 immediate reasons why someone would be referred to
by their network address: a) it provides UNIQUE identification as to
just *who* the person is (for example: if you are talking about Joe
Smith, there could be *many* Joe Smiths around the network, but only
1 Joe@Harvard), and b) if you're reading the message, and you're
lazy like I am, you don't want to have to look back at the headers
to find out just *which* Joe Smith the author was talking about!
- Andrew
- BANDY@MIT-AI
- BANDY@MIT-OZ@MIT-ML
(AI is pretty dead these days)
------------------------------
Date: 30 Aug 82 13:42:54-EDT (Mon)
From: Andrew Scott Beals <andrew.umcp-cs@UDel-Relay>
Subject: getting non-technical people to use computers
Take away their typewriters, scratchpads, calculators, file
cabinets, and 3x5s! Make 'em use the little beasties.
But, on a more sober note, show them that it *is* better/easier than
doing it by hand, and unless they are a technical fraidy cat,
they'll use it.
- Andy
- BANDY@MIT-AI
P.S. I guess I tend to be a bit radical at times...
------------------------------
End of HUMAN-NETS Digest
************************