perm filename HUMAN.MSG[1,LES]8 blob
sn#547948 filedate 1980-11-26 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00022 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00004 00002 GTE UNIT PLANS TO OFFER ELECTRONIC MAIL SERVICE
C00006 00003 Subject: LCD terminals
C00008 00004 Today's Topics: Commentary on Modems, Get together at NCC?
C00013 00005 Subject: Modem madness
C00025 00006 Subject: IBM speaks
C00029 00007 JDD@MIT-ML 05/01/80 01:35:41 Re: Remote-Access Games
C00035 00008 Subject: Large articles for FTP distribution
C00039 00009 Subject: Stock quotes etc. Via terminal
C00041 00010 << games >>
C00056 00011 Subject: local network announcement
C00062 00012 Subject: Random Ramblings ...
C00071 00013 Subject: Call for papers. IFIPS Symposium on Computer Message Systems
C00075 00014 Computer communication network bibliography
C00079 00015 SIRBU@MIT-MC 06/19/80 07:31:16 Re: French Teletel
C00082 00016 Subject: Dirty word filter
C00085 00017 Subject: An ETHERnet query from HUMAN-NETS
C00088 00018 Subject: More on Ethernet
C00089 00019 Subject: More on LEXIS full-text search system
C00093 00020 Subject: Screen editors for PDP-11
C00098 00021 Inventions in government-sponsored research
C00112 00022 Subject: Touch-Tone vs. Dial Pulses
C00123 ENDMK
C⊗;
GTE UNIT PLANS TO OFFER ELECTRONIC MAIL SERVICE
From the Wall St. Journal, April 3, 1980; page 32:
GTE Telenet, a unit of General Telephone and Electronics Corp. said
it plans to offer an electronic mail service starting July 14.
Called Telemail, the service will carry messages electronically
between various types of data terminals. In a filing with the
Federal Communications Commission, GTE Telenet said it plans to
charge 22 cents a minute during normal business hours and five
cents a minute otherwise, with prime-time discounts of as much
as 50% for high volume users.
Telemail will be different from current telex and facsimile services,
GTE Telenet said, in that it will be able to reach desktop and
portable terminals as well as centralized communications rooms. The
company estimated there are more than 1.5 million terminals already
in use that could be reached by Telemail.
------------------------------
Subject: LCD terminals
Date: 2 April 1980 23:47 est
From: SALawrence at MIT-Multics (STEVE at SAI-Prime)
Does anyone know how feasible it would be?
The most intriguing suggestion I've heard is to use an LCD panel on
the front of a CRT in place of the phosphor on current tubes; the
parts of the plate which were struck by electrons would turn dark.
The interesting property of the electron beam, from the point of view
of the LCD panel, would be its charge rather than the energy of the
electrons, and hence one could use far less energetic electrons, which
would make design of the deflection circuitry easier and cheaper. The
original suggestion I heard was to use this in an oscilliscope, where
it supposedly would let one deflect the beam far faster, and hence
display far faster events, with simpler deflection circuitry due to
the lower beam energy and electron velocity.
Today's Topics: Commentary on Modems, Get together at NCC?
HUMAN-NETS AM Digest Thursday, 24 April 1980 Volume 1 : Issue 63
----------------------------------------------------------------------
Date: 23 Apr 1980 0158-PST (Wednesday)
From: Lauren at UCLA-Security (Lauren Weinstein)
Subject: modems
There are several types of 1200 baud modems.
1) Vadic 1200/1200 -- full duplex 1200
2) Bell 212 1200/1200 -- full duplex 1200
3) split speed modems (usually 1200/150) 1200 out/150 in,
such as those from Prentiss and UDS
4) Prentiss 1200/1200
5) Vadic triple modem 1200/1200/300
The basic Vadic 1200 and Bell 212 are not compatible, the frequency
structure is inverted. Supposedly Vadic chose theirs to allow the use
of acoustic 1200 baud modems with fewer interference problems. The
Prentiss 1200/1200 is incompatible with both the 212 and Vadic. The
split speeds are only compatible among themselves. The Vadic triple
will operate with either
1) another Vadic
2) a Bell 212
3) an ordinary 300 baud type (103) modem
The bottom line with 1200/1200 baud modems is, cost. I believe Vadics
run about 3-4X the cost of good 103 type modems. Same goes for the
Bell types of course, if you calculate out equivalent costs. 1200
baud is the wave of the future, but there are several things to note:
1) there are alot more 300 baud ports out there than 1200 baud ports.
This applies to most computer centers, Telenet, Tymnet, GEnet,
etc. This will change over time, but cost will remain an
important factor until prices come WAY down. The issue of whether
a given computer center should use 1200 or 300 baud modems for
their remote users is one that arouses considerable emotion, and
no doubt we will see some of this in discussions of this topic on
HUMAN-NETS. There are those who claim that less than 1200 baud is
useless for real work, while others will say that 300 baud is
perfectly adequate with intelligent terminals and smart display
editors, and that the cost difference with the 300 baud units
allows more people to enjoy the benefits of remote access.
2) If you do not have really clean telephone connections, or if you
are not local to your dialup, watch out! The error rate on 1200
baud full-duplex connections is considerably higher than at 300
baud on other than "clean" circuits, and on some long-distance
circuits may become intolerable at times. Apparently there are
even worse problems with services like Execunet and SPRINT -- the
quality of connections is sometimes so bad that 1200 baud cannot
be used at all. Most successful use of 1200 baud has been with
very local users.
Will all modems be 1200 baud? Well, maybe eventually, but not for
quite some time. 300 baud still definitely has its place, especially
with intelligent terminals and terminal drivers.
--Lauren--
p.s. For a lot more detail and the names of manufacturers try DATAPRO
REPORTS ON DATA COMMUNICATIONS.
Subject: Modem madness
Date: 26 Apr 1980 1654-PST
From: Les Earnest <LES at SU-AI>
It is very hard to understand the popularity of the various 1200/1200
modems since they cost much more than the 1200/150 modems, offer no
better performance for terminal connections, and seem to be less
reliable. Why then do people buy them?
We have been buying and using the UDS split-speed modems since they
became available several years ago and find them quite reliable. A
local graduate student conspiracy recently developed a 1200/150 modem
that is Prentiss and UDS compatible and that comes as a kit for $75.
It plugs into an S-100 bus.
The 1200/1200 modems do make sense if you are trying to do file
transfers or other large data transfers both ways between two
computers. For this reason, we chose the Vadic 1200/1200 modem to
develop the first version of Dialnet, which offers Arpanet-like
services over dialup phone lines.
It is rumored that a certain major modem manufacturer is developing a
4800/150 modem with turnaround capability that works on dialup lines.
This could provide rather good terminal service as well as being
useful for file transfer work.
CBF@MIT-MC 04/27/80 06:08:40 Re: 1200 baud communication
Les Earnest asserted that 150/1200 baud modems cost "much" less than
1200/1200 modems. I cannot really speak to the $70 parts cost for a
homemade modem, we all know thats lots of things have cheap parts
cost if you're willing to build and debug them yourselves. I do know
UDS makes and sells good quality modems for a better then competitive
price. The UDS 150/1200 modem lists for $575 in my price list (a
little old, its prior to direct connect). A comparable Vadic 3455
lists for $825 (and is direct connect). Perhaps that is a lot
cheaper, but note that UDS's own 1200/1200 modem sells for just $600.
No one in the previous messages mentioned the UDS 12-12, which is the
price leader in 1200/1200 modems, unfortunately it is neither Vadic
nor Bell 212 compatible. The last two organizations I worked for also
each had access to Vadic's at quantity, educational or GSA prices
that brought the Vadic 3455 down to $700. The best UDS price break
brings the 150/1200 down to $530. These prices are all in a box with
power supply, please do not send in responses comparing card-only
prices.
Note also that the 1200 baud full duplex modems (Bell 212, Vadic 3400
and UDS 12-12) have better noise immunity than the 150/1200 baud
modems. The 150/1200 baud modems are all Bell 202 protocol which is
FSK (Frequency Shift Keying) whereas the 1200/1200 all use some form
of PSK (Phase Shift Keying) which makes better use of the available
bandwidth.
I will concede that 150/1200 operation has a moderate cost advantage.
I expect the cost-advantage to decrease in the future because while
the parts cost advantage of a 150/1200 over most 1200/1200 designs
is considerble, there will be more investment into cost-reduction of
1200/1200 designs. I understand the Bell 212 already contains some
custom LSI which Bell has liscenced to almost all the independent
vendors of 212 equivalents (with the notable exception of Vadic
probably because their designs are all 3400/212 hybrids with
considerable shared circutry).
The main question then becomes whether the cost diference is worth
keeping around 2 types of modems when one will do. Ie. I assert (as
Les has admitted) that there are applications which require 1200/1200
operation, though one can often get away with 150/1200 operation for
simple dialup use. When the price difference is $700 vs. $530 though,
I would go for increased standardizaiton as being more cost-effective
for an organziation in the long run. There are also the minor but
quite useful additions that come with modern modems, like local and
remote self-test. The considerations have to also include whether your
mostly talking to just yourself, the rest of the world or both. The
latter two cases are where the Vadic triple modems become a great buy.
Note that Telenet's and Tymshare's 1200 baud network ports around
the country are either Vadic or Bell 212. Many other commercial
installations as well. In fact the only great concentrations of
150/1200's I know of are in and around the Stanford community.
I should also point out that someone mentioned the Prentice 12-12's
in one of the earlier replies. Note that these are really split speed
hacks which merely run the recieve clock at 1200 baud so the host
terminal or host computer need not have split speed capablilty.
However, the aggregate data of the terminal to computer connection
must not exceed 150 baud. For some reason Prentice does not always
make this completey clear in their advertising. Of course this hack
means they cannot talk to anything but one of the same type at the
other end. The Bell 212, Vadic 3400 and UDS 12-12 are the only true
1200 baud full duplex protocols I know of. Of course, the Codex 24/24
(2400 baud full duplex) qualifies too.
------------------------------
Subject: IBM speaks
Date: 27 Apr 1980 0614-PDT
From: Scott at SRI-KL (Scott J. Kramer)
BC-SPEAK-04-27
By Adrian Berry, Daily Telegraph, London (Field News Service)
A machine that can listen to human speech and print what it
hears has been developed by American computer scientists. This is
a great step forward in technology and can bring nearer the day of
the ''intelligent machine.''
The machine, the IBM System 370, is programmed to understand a
vocabulary of 1,000 English words, Frederick Jelinek, of IBM, told a
recent meeting of the American Society of Automotive Engineers in
Houston.
''These results are an encouraging early step along an enormously
difficult path that some day may lead to computer recognition of
unlimited continuous speech, Jelinek said.
But the machine's performance is still primitive by human
standards, and it is easily confused by speech idiom. A sentence
beginning: ''Although the invention has been described...'' was
mistranslated as, ''All of the invention has been described ...''
It took 100 minutes to transcribe a spoken sentence that took a
human 30 seconds to utter, and anyone speaking to it had first to read
a series of 900 sentences into a microphone so that the machine could
get accustomed to his voice.
In a related development, a team of doctors at the University of
Wisconsin Medical school have found that computers can predict suicide
much more accurately than human doctors.
In a study of 63 mentally disturbed patients, doctors failed to
identify the three who afterward tried to kill themselves, while the
computer picked all of them.
Another major reported advantage in using computers for
psychiatry, with the machine asking ''yes'' and ''no'' questions, is
that a 40-minute interview costs only a few cents in computer time.
------------------------------
End of HUMAN-NETS Digest
************************
JDD@MIT-ML 05/01/80 01:35:41 Re: Remote-Access Games
Here is part of an article in this week's Chicago Reader:
"The brainchild of a bunch of frat brothers from Wabash College,
the Chicago-based GameMaster is a computerized playground that
can be accessed by anyone with a computer and a modem ...
"... GameMaster is the owner of a six-story ``mansion'' containing
over 30 electronically constructed ``rooms''; to get in on the
action, a player enters the foyer by dialing the GameMaster phone
number, then moves freely through the ``mansion'' by typing
directions into the computer. All of the games -- from chess and
bridge, to specially designed contests on sports themes and
military strategy -- are found in the basement. But games are only
part of what's going on here.
"While all the rooms aren't finished yet, the ultimate plan is to
make GameMaster an information center as well. Thus, a stop in the
kitchen allows the computer-user to call up a file of recipes
submitted by various subscribers; the newsroom will supply an
updated headline service on current events; in the mailroom, one
can check his ``mailbox'' ... for any messages left by other
GameMaster users; and the fourth-floor classroom will supply
programs for the study of math, history, foreign languages, and
other subjects.
"GameMaster's most significant innovation may be the interaction
among its various guests. The GameMaster computer makes a worthy
opponent in any of its games, but it's also content to serve as a
referee between two human opponents: as you type your direction
into your computer in Chicago, GameMaster can instantly relay them
to the screen of a player in, say, Waukegan, allowing a subscriber
to compete with someone he's never met -- except electronically.
And how does one meet such opponents? Why, in the parlor, of
course: a new player is advised to stop in this ``room'', type in
a salutation, and wait for GameMaster to send him the
introductions of other players who've stopped in to see who's
around. If two of you agree on a game, it's off to the library to
peruse the rules, then down to the appropriate game room.
Everything ``spoken'' in the parlor can be ``heard'' by everyone
in there, but more private discussions can be held in one of the
conference rooms -- electronically ``locked'' to the public.
"... GameMaster finally opened shop last fall, and now has 40
subscribers (who pay a $50 initiation fee and about $2.75 per hour
of computer time); [the owner] figures the system needs about 100
members to break even. (There's also an ``open house'' the first
Friday of each month, whereby nonsubscribers can try out the
gaming by calling [312] 726-8260.) ..."
Related questions: what types of games would be best to play over a
WorldNet? What effect does the organization of a WorldNet (e.g., a
centralized host vs. many small computers talking among themselves)
have on the types of games that are possible/feasible? Are there
reasonable games that 1E6 people could play at a time? Is the
Adventure paradigm a useful one in structuring access to network
resources? Will the currently many small innovators be crushed when
the big companies decide to enter the field?
Subject: Large articles for FTP distribution
Date: 5 May 1980 0314-EDT
From: Roger D. Duffey, II <DUFFEY at MIT-AI>
I have two articles for HUMAN-NETS that are too large for digest
distribution. They are described below. For your convenience the
material has been established in files at four different sites.
Please note that accounts are not necessary to TYPE or FTP the files
from SU-AI. The files will remain available for 7 days. Afterwards
their contents will be archived with the other HUMAN-NETS materials.
Questions or problems in obtaining copies of the files can be directed
to me. Also thanks go to Richard Brodie, Richard Lamson, and Don Woods
for their time and file space.
Material on the Heathkit H19 terminal:
This is the material I mentioned in HNT V1 #72. The file contains:
1. an evaluation of the terminal prepared by Charles Frankston
<CBF at MIT-MC> ( with substantial contributions by Winston
Edmond <Edmond at BBN-TENEXE> ),
2. the messages of wide interest sent to INFO-TERMS about the H19.
The file is organized as a sequence of messages (unfortunately still
in reverse chronological order). The messages have been edited for a
70 space line.
Site File Name
MIT-AI DUFFEY;HUMNET H19
MIT-Multics >udd>PDO>Lamson>human-nets>human-nets.h19.text
PARC-MAXC [MAXC]<Brodie>HUMNET-H19.TXT
SU-AI HUMNET.H19[T,DON]
Article from the Boston Globe on Electronic Mail Systems:
This article discusses commercial efforts directed towards Electronic
Mail Systems. The article mentions GTE "Telemail", CCA "Comet", BBN's
"Hermes", Tymnet's "On-Tyme-II", Wang's "Mailway", among others.
[ Note this article has already been distributed to MSGGROUP. ]
Site File Name
MIT-AI DUFFEY;HUMNET MAILER
MIT-Multics >udd>PDO>Lamson>human-nets>human-nets.mailer.text
PARC-MAXC [MAXC]<Brodie>HUMNET-MAILER.TXT
SU-AI HUMNET.MAI[T,DON]
Subject: Stock quotes etc. Via terminal
Date: 5 MAY 1980 0652-PDT
From: RAHE at USC-ECL
There are at least two other sources for stock market quotes other
than The Source: Dow-Jones supplies a stock/bond/futures quote system
with local numbers in many large cities. The price is rather high:
$3 per connect plus $10/hr. The typical 1 or 2 stock quote will then
cost about $3.25 or so. For 300 bps terminals max.
Western Union also has a service which lets you get news, sports
finance, stock market and in addition weather, silver and gold
futures, grain and livestock prices and, in season, ski reports!
You can also have the summaries(except ski and stock) sent to your
TWX/TELEX terminal at times you specify. Don't have a cost figure
tho.
<< games >>
∂14-May-80 0038 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V1 #83
Date: 14 MAY 1980 0311-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V1 #83
To: (@FILE [DSK:AIRAN;DUFFEY HNTFMC]) at MIT-ML
HUMAN-NETS AM Digest Wednesday, 14 May 1980 Volume 1 : Issue 83
Today's Topics: Politics and Nets, Games 1E6 People Play,
3 New Computer Network Offerings, Reply on Sexism
----------------------------------------------------------------------
Date: 13 MAY 1980 1747-PDT
From: MACKENZIE at USC-ECL
Subject: US GOV Network
Have you ever known any major candidate in these days of media
people that would make a position statement on anything that meant
anything? Read your voter's pamphlet!
Chuckle,
Kevin
------------------------------
SIRBU@MIT-MC 05/13/80 22:55:11
Regarding nets and politics: There is already a database of the
status of Federal legislation. It is maintained by the Library of
Congress and you can walk into a quiet room in the LoC and check
the current status of any bill. A number of states are installing
similar systems (including Massachusetts).
As for files of position statements, the White House is currently
construciting such a database for the president. Maybe then he won't
contradict himself so often.
------------------------------
Date: 13 May 1980 1005-EDT
From: JHENDLER at BBN-TENEXA
Subject: Games 1,000,000 people play
There was some discussion lately about the games on the PLATO (CDC)
system. As the resident expert on playing with TO games, let me
comment.
1) Empire exists on PLATO, There are four competing teams (Feds,
Orions, Klingons, and Romulans) and most of the features discussed
in prior messages. People come and go as they please, but there
contributions to the gameremain, the object is for one team to
conquer the universe, which happens roughly once a month or so. I,
myself, haven't played this one much, but many people swear by it,
and say the more the merrier.
2) There are various D&D type games on PLATO, but only one which
is multi-terminal, this is called "The Mines of Moria." Unlike the
standard D&D one can travel either with a group or not , as one
pleases, and when one leaves the game his/her character is left in
the state it was when he/she left. The other players in a group can
continue on, or all can agree to meet again at a later time. One can
travel with up to 9 other players, and move through "rooms" each of
which is a 6x6 grid (walls, doors, etc). Players can give each other
food, water, or weapons/magic items.
The dungeon is structured by having a city (top-level) weith magic
stores, weapon shop, water houses, and supply stores. Connected to the
city is the wilderness, with connections to four terrains, each of
which has 60 levels. As one goes deeper, he/she meets more and tougher
monsters, but gets bigger treasures, etc. One gets stronger as one
progresses. Other features include camps (where food and H2O can be
stashed), water holes (which may or maynot be poisoned) and guilds.
Anyone wants more info let me know
Jim (Jhendler@BBN)
p.s. both of these games could probably be extended to 1E6 players,
but I'd hate to see the load average on those machines....
------------------------------
Date: 13 May 1980 09:41 PDT
From: Weissman at PARC-MAXC
Subject: Megaplayer Games
Referring back to the message (in H-N v1#70) that, I believe, started
the discussion on megaplayer games, yesterday I received my "personal
invitation to enter the GameMaster system"; I had gotten on during
their "open house" & left my address for their brochure. I chatted for
a bit in the "parlor" of the "house" with a member of the GM staff,
who asked me how I had heard about the system. It turns out these guys
have never heard of the ArpaNet -- the guy I talked to, at least, was
boggled by the idea of a nationwide mailing system.
As for the games themselves, most are pretty ordinary fare: arcade,
board, and card games, etc. plus some non-games: message systems,
teaching systems, and the like. But the most interesting in terms
of the current Human-Nets discussion, D&D, etc., is their
partially-implemented fantasy game. This game is in three parts; I
quote from the GM brochure:
GARNFORM'S MAZE -- This old codger is the keeper of the CATACOMBS but
before he lets you in, you must find the key that he has hidden. Once
found, the key unlocks the door to the CATACOMBS. (On-line)
CATACOMBS -- All those entering this 4-level-deep cavern to so as
warriors. Each level provides a clue to descending to the next layer.
Creatures, curses, and treasure are scattered about, but your main
objective is to procure the passport that will get you transported to
the city of RASKAR. (Forthcoming in May)
CITY OF RASKAR -- This is the starting point of the adventure of
a lifetime. Upon your arrival, you will be joining a perpetual
interaction of machine and human characters. Those first few will
be true pioneers of this new land and most likely will have
advantages over the latecomers. (June)
The brochure doesn't give any more details on the Raskar game, or
specifics on how the "perpetual interaction" is to be maintained,
e.g., what happens if you log out. For more info, you can write to
GameMaster, PO Box 1483, Evanston IL, 60204 or dial-in to (312)
726-8260 during their next open house, April 4.
------------------------------
Date: 13 May 1980 13:03 PDT
From: JimDay.ES at PARC-MAXC
Subject: Nestar Nets
Nestar Systems plans to provide hardware capable of controlling
networks of Apple II computers. Intended for use by large
corporations, a typical network of 50 to 60 Apples would cost
between $50,000 and $100,000. Nestar is looking for foreign and
domestic computer retailers willing and able to handle the needs
of such a market. Applications software will not be provided by
Nestar, they prefer to leave software development up to the end
user or independent software vendors. Their Cluster/One Model A
can link as many as 65 Apple computers into a net providing as
much as 33M bytes of shared disk storage. Each Apple will be able
to operate independently, with access to files stored on the net.
Another Nestar product is the Cluster/One Model One, intended
primarily for educational users. It is capable of supporting nets of
Apple, PET, or TRS-80 machines, but without the data file capability
of the Model A.
For more information, write to NESTAR SYSTEMS, INC.
430 Sherman Avenue
Palo Alto, CA 94306
or call (415) 327-0125
------------------------------
Date: 13 May 1980 16:20 PDT
From: JimDay.ES at PARC-MAXC
Subject: IBX/S40
Exxon affiliate Intecom, Inc. has announced its new Integrated
Business Exchange (IBX/S40), designed to handle both synchronous
and asynchronous data in the 110-56K bit/sec range. No modem is
required, and voice traffic can be handled simultaneously over
standard telephone lines. This and other Intecom controllers permit
communications between data processing, word processing, electronic
mail, video, and voice equipment. More than just a data switch, the
IBX/S40 translates a wde range of protocols and converts data
formats, supporting the packet services of Tymnet, AT&T, GTE, RCA,
Western Union, etc.
The IBX incorporates a nonblocking switching matrix and multiple
processors, each capable of storing as much as 4M data bytes and
coordinating the operation of up to 16 distributed subsystems.
Using Zilog Z-80 microprocessors, each subsystem provides up to
256 I/O ports for communication with peripheral devices.
Other Intecom devices digitally encode voice inputs. The IBX can
transmit multiplexed digitized voice and data via RS-232C or RS-449
interfaces through ordinary twisted pairs over distances of a mile
or more. The IBX design also provides hardware redundancy and remote
diagnostic capabilities for reliability. For more information, write
INTECOM, INC.
16509 Addison Road
Dallas, TX 75240
------------------------------
ZRM@MIT-MC 05/13/80 11:55:48
Re: Sexism [ see HNT V1 #75,77,78 for related messages. -- RDD ]
I was talking about the type of people who play games like the
Mass. state lottery and other gambling games of hopeless odds, and the
people who watch game shows. Unfortunately, this group of people tend
to be working/ middle class white suburban housewives. I was
commenting on the state of society and not supporting the brainwashing
of women -- in fact I was lamenting it. Computer games, particularly
those that use the raw data processing power to manage mass
participation and neglect the ability of computers to stimulate minds
with intellectual challenges, are dangerous. The prospect of further
anesthetizing the minds of the valium addicts is frightening. Who
needs Stepford when we may have to deal with organic automata?
--zig
Addendum: This reply to the comments about my supposedly sexist
flamage was delayed by my being out of town and due to the fact that
Roger returned my mail with a few very helpful suggestions for which I
am grateful.
------------------------------
End of HUMAN-NETS Digest
************************
Subject: local network announcement
Date: 13 May 1980 16:27 PDT
From: Shoch at PARC-MAXC
Digital Equipment Corporation Intel Corporation
Maynard, Massachusetts Santa Clara, California
Xerox Corporation
Stamford, Connecticut
"DEC, INTEL AND XEROX DEVELOP NETWORK SPECS"
LOS ANGELES, May 13 -- Digital Equipment Corporation, Intel
Corporation and Xerox Corporation have announced a joint project to
develop electrical, logical and protocol specifications for a local
area communications network which the companies propose to use for
many of their future products.
The local area network consists of a coaxial cable and communications
transceivers designed to link different kinds of computers, computer
peripherals, data terminals and office equpiment located in a building
or in a complex of closely grouped buildings. Each device connected to
the cable will contain a control element allowing it to communicate on
the cable through its transceiver. The cooperative effort by the three
companies, begun last spring, will result in specifications expected
to be published in the third quarter of 1980.
Xerox is providing the basic local area network design represented in
the Ethernet network capability, announced by the company in December,
1979. Experimental Ethernet networks of several hundred stations have
been used over a period of five years in several Xerox sites. Digital
is providing system design expertise in the area of communication
trnasceivers and microcomputer, minicomputer and mainframe computer
networks. Intel is providing expertise in the partitioning of complex
communications functions into microcomputer systems and VLSI
components.
Since one of the objectives of this joint project is to provide
communications compatibility amoung users of a broad range of
computers, computer peripherals, information systems and office
products, use of the specifications by other corporations and
organizations will be encouraged. In particular, members of the
joint project will cooperate with any standards organizations
expressing interest in adopting the specifications. The three
companies agree that providing for the low-cost interconnection
of information processing systems and products on a local
communications network will be a vital element in the evolution
of future information systems.
Although testing is incomplete, several specifications have already
been established. For example, the network will operate at 10 million
bits per second, and a typical configuration will consist of one or
more segments of coaxial cable each of which can be up to 500 meters
in length.
In order to promote widespread use of this communications capability,
Xerox, which holds patents on the basic Ethernet techniques, said it
will make patent licenses available to interested parties including
those who wish to manufacture compatible products. Several computer,
office equipment and semiconductor manufacturers have expressed
interest in manufacturing compatible systems or components.
Contact: Dick Berube, Digital, Maynard, Mass (617) 493-3046
Jim Jarrett, Intel, Santa Clara, Calif. (408) 987-5090
Sandy Lanzarotta, Xerox, Los Angeles, Calif.
(213) 679-4511, ext. 7181
------------------------------
Subject: Random Ramblings ...
Date: 14 MAY 1980 0143-EDT
From: JDD at MIT-ML (John DeTreville)
The imagery underlying some of the recent discussion is fantastic!
We have built new bodies for our old minds, and now we build new
houses for them to live in. The new bodies are our net-hopping
selves: sexless, raceless ghosts of our own making, zipping across
a landscape swept clean by the packetized winds of change; the new
houses are the networked rooms which they will inhabit for a while,
interacting with the others they may find there. If we use magical
rules to structure this space, this is only because magic is the way
we'd like the world to work, and now we have our own world to try it
with. The mythic possibilities are staggering....
But, back to work. Here are a few semi-related points.
Games are one point of view to help us understand a WorldNet.
Consider another: the publishing industry. We can imagine, for
example, electronic magazines, published at a central source and
distributed over the net to the subscribers at regular intervals
(let's assume for the moment that the subscribers have a home
computer with adequate storage capability). Now: we can either send
one copy at a time, n times, which could cost quite a lot, or we
could broadcast the magazine, sending it out to local distribution
centers and then to the subscribers, which could be cheaper [it
might be cheaper yet to broadcast directly by satellite]. What do
the economics look like for these various approaches? The broadcast
approach, although cheaper, might lose out on flexibiity. In fact,
if we didn't publish at all, but simply kept one copy on-line at
some central place, we could have maximum flexibility at maximum
accessing cost (we could, for example, allow readers to make
annotations on the articles, which could immediately appear as
Letters-to-the-Editor-style footnotes).
How, then, should things be done? And what about copyrights? Right
now, I can subscribe to a magazine, photocopy it, and give the
copies to my friends, but that takes a lot of effort on my part: it
would be much easier (e.g., automated) with the WorldNet. Or would
the economies of scale (if they continue to exist) still make it
cheaper for my friends to have their own subscriptions? Anyway, all
this is one scenario: what are others? Will the relatively high cost
of artwork keep it out of net magazines? How will libraries be
affected?
Publishing is one thing that we can expect to see continue in a
maybe-recognizable, maybe-not form under a WorldNet. Let's
semi-consider another: shopping centers. These would be a group of
net stores (net addresses, either automated or semi-automated), with
linking information to make them "close" to each other. Sounds
reasonable. But: shopping centers were, the story goes, a by-product
of the automobile, which optimized some actions at the expense of
others. WorldNet will presumably optimize a different mix of items;
will structures proven in the old system remain true in the new one?
What new, unexpected organizations will rise up out of the new? Will
anything remain the same? Or is all of this blue-sky theorizing
totally irrelevant, and will things remain exactly the same?
In the shorter term: It seems plausible for many reasons that the
Bell System could/will serve as the basis for a WorldNet (or at
least an AmericaNet). What should the Bell System start to do now to
bring this about as expediently as possible? Why should it? (One
reason it shouldn't is obsolescence: how can a nationwide system be
planned when the technology is so volatile? A 50-year-old telephone
hooks up to the telephone network just fine today: engineering
decisions made today could haunt us for another 50 years.)
Finally: It seems plausible that a WorldNet could remake society.
Question: is this good or bad? We can all think of tons of great
things one could do with a WorldNet, but what could a WorldNet do
to you? There were plenty of high hopes for commercial television
in its early days, but these hopes seem totally unrelated to what
it has become.
The Journal of Nutritional Academy reports that third grade
students in a Connecticut grammar school class were asked to
spell the word "relief". More than half the kids spelled it
"R-O-L-A-I-D-S".
As a recycled example, you could use a Worldet to access your
psychiatrist, but isn't that incredibly dehumanizing? On the
other hand:
Researchers at Britain's National Physical Laboratory report
that their patients find it easier to talk to computers than to
therapists. The Laboratory has installed a computer named Mickie
to question patients suffering from alcoholism, and they report
that patients are far more likely to admit high alcohol
consumption to the machine than they are to an inquiring doctor.
How does Mickie succeed where humans fail? The researchers believe
it's because the computer merely asks and receives an answer, and
doesn't make the moral judgements a human therapist might when
questioning patients. England's Department of Health is reportedly
so impressed by Mickie's performance, it has decided to buy five
more computers. [Both excerpts from Wet Magazine.]
Marshall McLuhan has espoused the idea that "the medium is the
message": that the form of communication changes the communicators,
and, in doing so, ultimately overshadows any abstract content. How
can we understand the WorldNet in this context? We've already talked
a lot about physical effects (e.g., radiation): what about the
mental ones? What about the sociological ones?
Subject: Call for papers. IFIPS Symposium on Computer Message Systems
Date: 29 May 80 15:54-EDT (Thu)
From: Farber at UDel-EE
Reply-to: Farber at Rand-Unix
Call for Papers
IFIP TC-6 International Symposium on Computer Message Systems
Montreal, Canada
6-8 April 1981
Computer mediated messaging is a rapidly emerging new service
area on the international scene. Messages may be processed,
stored, and transmitted bewteen users potentially within the
jurisdiction of separate carriers, computer systems, and/or
computer networks. This will be an international symposium aimed
at promoting interchange of information and discussion on
technical, economical, and political issues. Papers are desired
from all topic areas.
Papers must be received by 30 Sept 1980 for consideration.
Papers will be reviewed, and authors will be notified by
Nov 1980 regarding acceptance.
The Program committee members for North America are:
David Farber - USA
Ronald Uhlig - Canada
Complete copies (not abstracts), in 4 copies should be sent to:
Ronald P. Uhlig, BNR Ltd. Dept 3D20, Box 3511, Station C, Ottawa,
Canada KIY 4H7
Computer communication network bibliography
Date: 10 JUN 1980 0821-PDT
From: Hathaway at AMES-67
Subject: Results of textbook information request
This is a report on the results of my request for information on
textbooks for a course on "computer communication networks," with
emphasis on software protocols and social applications.
The request went out to MSGGROUP and HUMAN-NETS on June 5. I have so
far received 18 responses, with 14 making specific recommendations on
books and 6 requesting that I prepare a summary of what I hear (which
is what you are reading ...).
The recommended books are:
Computer Networks and Their Protocols
Davies, Barber, Price, and Solomonides; Wiley, 1979
(7 recommendations)
A Practical View of Computer Communications Protocols
McQuillan and Cerf, IEEE Tutorial Series, 1978
(3 recommendations)
The Network Nation
Starr Roxanne Hiltz and Murray Turoff
(2 recommendations)
Advances in Computer Communications and Networking
Wesley Chu, 1979
Communications Technology and Social Policy
Gerbner, George, Larry Gross, and William Melody (eds.)
Computer-Communication Network Design and Analysis
Mischa Schwartz, Prentice Hall, 1977
Queueing Systems Computer Applications (Vol II)
L. Kleinrock, 1976
Technical Aspects of Data Communication
Digital Press
Telecommunication Transmission Handbook
Roger Freeman
Toward Paperless Information Systems
F. W. Lancaster
"Doll's book on computer communications"
"All of James Martin's books"
"one forthcoming later this year by Tannenbaum, Prentice Hall"
Arpanet protocol handbook
DEC DNA architecture series
IBM SDLC series
The following journals were also recommended:
IEEE Transactions on Communications, April 1980
On-Line Review (has published three bibliographies)
Proceedings of the IEEE, November 1978
All in all, a very interesting and valuable exercise. Thank you to
all who participated.
Wayne H.
------------------------------
SIRBU@MIT-MC 06/19/80 07:31:16 Re: French Teletel
Last week I saw a demonstration of the French Teletel system, their
version of viewdata/Prestel. At the end of this year they will begin
a demonstration with 3000 subscribers who will receive free hardware
and offered a variety of services at $.025/minute connect time.
Since January 1 all TVs sold in France must have a plug which
gives access to the composite video input, or direct access to the
Red-Blue-Green. Thus it is easy to design a box to plug into the
TV without worrying about RF modulators.
The Teletel box (current price 2000 francs = $500) is 18 x 10 x 2
and is designed to lie flat and sit under the TV set. The detatched
keyboard is battery powered and communicates to the box by infra red.
Keyboard has full alphabet but is hand calculator quality.
The French are pushing transaction services, not just database
retrieval. The Teletel computer will be linked via the French packet
network (Transpac) to Air France reservation computers, banks, catalog
retailers, so that users can make reservations or order goods. There
will be a mailbox service offering prepared "cards" (happy birthday
written out in block letters using color graphic symbols) and free
text. They are pushing mail to advertisers for direct advertising.
The French PTT expects terminal prices to drop to 300 francs ($75)
in volume production.
The system supports both tree-structured retrieval as with Prestel,
and use of named commands.
Speed is 1200/75. Limitation is quality of local loop wiring,
particularly in central cities like Paris where wiring is old.
All in all, the French seemed to have put the pieces together
somewhat better than the British (or the Americans such as Source
and CompuServe). It should be interesting to watch how they do.
Subject: Dirty word filter
Date: 10 July 1980 18:28 edt
From: Schauble.Multics at MIT-Multics
The University of Maryland used to have a filter on their RFOR
Fortran compiler. Seems that they had a problem with students using
nasty words as variable names. It was removed when a professor in
the agriculture department had a program rejected. He was doing a
study on chickens, and his program had variables for "cocks" and
"hens". He raised the roof.
Subject: Censorship
Date: 14 July 1980 0509-EDT (Monday)
From: Brian Reid at CMU-10A
In the early 1950's censorship was common in American publishing.
Today's gag phrase "banned in Boston" was no joke back then. One
of the most erotic novels in English, Vladimir Nabokov's classic
@i[Lolita] published in 1955, was carefully written to evade the
censors by using a harmless vocabulary. The trick worked: the book
was not banned, the vocabulary passed muster, and the book lives
on as a libidinous classic.
As an aside, I should confess that I am the author of the
University of Maryland's dirty-word filter that Paul Schauble
mentioned a few days ago; it was written in 1967, before I'd read
@i[Lolita]. I didn't even bother with "Hell" and "Damn"; in 1967
those were prepositions.
Brian
Subject: An ETHERnet query from HUMAN-NETS
Date: 24 OCT 1980 1313-PDT
From: SHOCH at PARC-MAXC
1. The Ethernet spec is not available on line.
2. The spec is available from DEC (in Tewksbury), Intel (in Santa
Clara), or Xerox (in Palo Alto) -- whoever you normally like to
deal with.
In particular, to get it from Xerox drop a note to:
Bob Printis
Xerox
3333 Coyote Hill Road
Palo Alto, California 94304
John Shoch
Subject: More on Ethernet
Date: 29 October 1980 1321-EST (Wednesday)
From: George Marshall
Additional information in response to the request for Ethernet
information on HUMAN-NETS:
As John Shoch mentioned, specs are now available from all three
companies involved. If you would like receive a copy from Intel,
and thus be put on the mailing list for its future announcements
of component and/or system- level Ethernet products, send your
request to:
Intel Corporation
Ethernet Literature Department
3065 Bowers Avenue
Santa Clara, CA 95051
Subject: More on LEXIS full-text search system
Date: 5 Nov 1980 at 1101-CST
From: amsler at UTEXAS-11
I have been tracking down further information on LEXIS, the Le-
gal Document full-text system. LEXIS is marketed by Mead Data
which is a subsidiary of the Mead Paper Co. (guess what the
phrase "paperless society" does to a Paper Company! They invest
in paperless information systems). LEXIS is a booming success!
It currently has 2 billion words of text on-line!!! As far as I
have determined they acquire most of their material in machine-
readable form already. LEXIS charges $125/hour with a $500/month
charge as well. At Univ. of Texas it is accessed via a local 1200
Baud dial telephone line.
There is apparently a second massive full-text system, called
WESTLAW, which was started by the West Publishing Co. (a Legal
publisher who already HAD machine-readable data from their photo-
typesetting operations). They ALSO have 2 billion words (English
words) of text on-line.
What all this is suggesting is that massive full-text databases
are now commercially possible IF the price is right. Mead Data
sold their full-text indexing system to the Boston Globe for use
in managing their newspaper morgue. The question remains where
will "electronic text" develop next. Mead has a NEXIS database
which deals with news service data (e.g. contains Newsweek maga-
zine, etc.), but I know little more about that.
So we have several pieces of the puzzle falling into place.
Text will be available in machine-readable form from publishers.
Either the publishers or someone who makes arrangements with the
publishers will gain access to the machine-readable data and be
evaluating the commercial prospects for selling access to it.
Once a market is determined (Laywers read a lot and have lots of
money and NEED information FAST (e.g. during a trial)) then an
on-line database will develop. Once the database proves it is
successful competitors will enter the field-- probably publishers
who were too timid to launch the project themselves.
So... who reads a lot and has lots of money? Stock Market
Analysts? Doctors? R&D Corp.'s, ??? the general public will have
to wait it would seem, although this may be more a matter of HOW
TO GET THE TEXT INTO THEIR HOME than their lack of wealth -- they
may be only average income ... but there are a LOT of them.
Subject: Screen editors for PDP-11
Date: 21 NOV 1980 0635-EST
From: FREND at MIT-AI (Michael A. Bloom)
There are a number of editors available for such use.
1) VTEDIT
A macro written in TECO that runs incredibly slowly. This is due to
a number of things. Vtedit has a predecessor called VT52.TEC. Unlike
the latter, VTEDIT manufactures a string equivalent to a label that
it will jump to for just about every command you type in. After
manufacturing this label (tag), it then procedes to search from the
beginning of the macro until it finds a tag that looks like the tag
it has built. If you have a recent release of TECO-11, you can fix
a large part of the speed problem tho, by replacing the jumps with
a newer flow-control construct. (and then replace the tag building
with simple conditionals). I'd guess at a speed increase factor in
the hundreds if you remove all explicit jumps.
VTEDIT can be made to run with non DEC terminals by changing some
escape sequence constant definitions in a file called CRTRUB.MAC this
is then reassembled, and linked with several other object modules
included in the TECO distribution. Unfortunately, the algorithm's
that control the screen come in the preassembled SCREEN.OBJ, which
does not know from line insert/delete nor char insert/delete. I
would LOVE to get hold of a copy of SCREEN.MAC (machine readable,
if anyone on this list has a copy).
2) KED and K52
These are DEC "keypad editors". I have never used them, but John
Wedel at the JPL Robotics Lab, uses KED all the time and is very
satisfied.
3) Amethyst
This is an editor which claims to be "a subset of MIT's EMACS".
It does not have TECO minibuffers, but it is written in C, and the
additional command parser to get a subset of TECO as well should
not be too difficult to add. I plan to try to get a copy and do
this for the 11 at my school.
4) TORES
Written by JIS, this is a screen editor that uses two UNIX processes,
one for editing and the other for screen handling. If you can use
pseudo keyboards on your RSTS system, you could probably start out
by just logging in another job to handle it. (or you could create
a runtime system to spawn the necessary job, and use SEND/receive
instead of the precious PK) It probably wont do to have too many
users using this one under RSTS/E. Oh yes, you'll also need a C
Compiler (I've heard there is one for RSTS).
5) TINTE
This is a text editor written in PALX (somewhat like macro-11, but
requires some editing to be assemblable by macro-- mostly straight
char string conversion ). It purports to be somewhat like Fine (I
have no idea) and contains conditional assembly to the extent that
not only can you assemble it to run on a number of different system
configs, but you can select whether all messages printed will be in
english or will be in french! Source to TINTE was on line last I
looked, but I dont remember who's directory it was in.
Inventions in government-sponsored research
∂24-Nov-80 1439 JJK
To: LTP, LES
Luis, Les -- I think the following exchange will be of interest to you.
It comes from a new Arpanet-wide mailing list of people interested in
computers and law. -- Jonathan
-----------
From: HEDRICK at RUTGERS
I would be interested in is some reasonably authoritative opinions on
a few legal issues. The most pressing are
- is it true that Federal support for a project means that
the public has some rights to the results. I had always
believed this, but several projects seem to have
resulted in software that somehow ends up being owned
by a company formed by the researcher. I would like
to know what options we really have with respect to our
own work, and what recourse we have towards researchers
elsewhere who use our tax dollars to create companies
for themselves. Currently we have no policy at Rutgers
about who owns software, whether it can be copyrighted,
and by whom, etc.
- if we have a project that believes it has a marketable
software product, what should we do?
- are there any laws relating to privacy of students' files on
an instructional system?
∂22-Nov-80 1204 Zellich at OFFICE-3 Federal support for a project...
Date: 22 Nov 1980 1201-PST
From: Zellich at OFFICE-3
Subject: Federal support for a project...
To: info-law at MIT-AI
In response to the question asked by Hedrick at RUTGERS:
It is a general (and almost mandatory) practice to require government
rights to any software developed under contract. In at least some cases,
however, a clause gets written in that gives the developer all COMMERCIAL
rights to the software - in other words, the government can redistribute
the contracted-for software only to other government agencies. I am
reasonably sure that contracts can also be written that give only the
specific contracting agency rights to use the software -- this is an
extremely poor procurement practice, and can get some VERY sharp questions
asked if it is noticed by GSA, GAO, etc.
-Rich
-------
∂22-Nov-80 1618 Michael.Shamos at CMU-10A Rights in Federally-financed software (long message)
Date: 22 November 1980 1908-EST (Saturday)
From: Michael.Shamos at CMU-10A
To: info-law at MIT-AI
Subject: Rights in Federally-financed software (long message)
Message-Id: <22Nov80 190814 MS20@CMU-10A>
In reponse to HEDRICK@RUTGERS:
Regarding the implications of government "support" for work done
at universities and non-profit research laboratories, a clear
distinction must be drawn between two cases:
I. The government has directly contracted for a particular item of
research or software. The subject matter of a government contract is
defined in its "Scope of Work" section.
II. Software development has gone on "at" an institution which receives
government funding in one way or another. The word "at" is in parentheses
because it is not always the case that work done by an employee of an
institution is necessarily done "at" that institution.
As far as case I, it is axiomatic that a contract only covers work
that is within its scope. For example, if the government contracts for
MIT to write a multiprocessing operating system and someone on the project
creates a video game having fantastic commercial potential, the game is
not within the scope of the contract and has not been purchased by the
government. (The rights of the developer are a matter between him and
his institution. He may get rich or he may get disciplined for making
unauthorized use of the equipment.) The rights of the government are
involved only to the extent of its contractual relations with the
institution.
At this point it should be mentioned that the rights of the government
are totally different from the rights of the PUBLIC. Many people assume
that they can obtain copies of anything owned by the government for just
a nominal distribution charge. This is absolutely false. The Freedom
of Information Act applies only to the operating records of government
agencies and is designed as a protection against federal mismanagement or
violations of civil rights. It is not a mechanism whereby intermeddlers
can obtain for nothing the fruits of intellectual property of others.
The leading case in this area is Systems Development Corporation v.
Mathews, 542 F.2d 1116 (1976). The facts are quite simple.
SDC applied to the National Library of Medicine under the FOIA to obtain
tapes of the MEDLARS medical information system. At the time, NLM,
which is a government agency (part of the National Institutes of Health),
was charging an annual subscription fee of $50,000 for the system. SDC
wanted to pay only the reasonable distribution cost provided by the FOIA.
Held, SDC was not entitled to the tapes, notwithstanding the fact that the
medical databases were assembled at a huge cost which was shouldered
entirely by the government. Conclusion: the fact that the GOVERNMENT
buys or pays for something does not entitle the PUBLIC to it.
The above principle is embodied in standard government contracts.
Even those which grant the government "unlimited rights" usually contain
a clause immediately granting back to the developer all patents and
copyrights that are not inconsistent with the rights of the government.
Thus, third parties have no right to use the discoveries or to insist on
receiving distribution copies.
Why is this fair? Why isn't the public entitled to what the government
has paid for? The answer is that it is federal policy to encourage
invention. Many potential inventors are not employed by the government
(a different situation) but work on federally-funded projects. To assert
that these people must give up all rights in what they do would be destructive
to their creative spirit. The fact that many inventors never bother to
assert their rights does not mean that they have none.
After this discussion, it should be clear that the public has even less
of a claim to work that is not directly, but only indirectly, financed by
the government. A typical situation is the one in which the government
pays a substantial fraction of the overhead in a university's budget
through research contracts. The money is used to buy a computer which some
graduate student then uses to develop software which has commercial
potential. I have heard people express the opinion that such activity should
be prohibited, the argument apparently being that nobody should get a
"free ride" at public expense. But one of the principal goals of
government funding of university research is to advance the state of US
technology. A large number of potential inventors would lose their
innovative zeal if they knew that nothing they did at a university could
be developed into a marketable product. The fact that someone may profit
from bringing technology to the marketplace is not regarded by our society
as a bad thing, as it recognized that some reward must be given for the
entreprenurial risk involved. It is naive to think that a piece of software
produced by a university author can be sold commercially over the counter.
Very often huge sums need to be spent to make a product suitable for
transfer to industry. SofTech estimates it has spent $900,000 developing
UCSD Pascal, and I can tell you from personal knowledge that about
$100,000 has gone into Scribe. (Scribe, incidentally, did not fall within
the scope of any government contract. I have no information on UCSD Pascal.)
There is a good deal more to bringing an invention to fruition than
the expenditure of machine cycles. Let us even assume that $1,000,000 in
chargeable computer time has been used to produce, say, a stock market
analysis program. What is the value of the human cycles that were necessarily
used during the development and for which the public has NOT paid? To
assert that the public should own or have access to the resulting system
is to ignore the value of the inventor's vision, creative energy and
toil. Government contracts, the patent and copyright laws try to strike a
balance between the rights of the various parties. The present situation
in the US is that creative people receive significant statutory protection.
In any event, whatever one's beliefs about the propriety of marketing
software developed in a university setting may be, the law and public
policy of the matter are clear -- it is encouraged, not discouraged.
Subject: Touch-Tone vs. Dial Pulses
Date: 24 Nov 1980 0348-PST (Monday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
There are basically two sides to the issue of "where is touch-tone
used, and to what end?"
1) The subscriber must tell the exchange the number he/she wishes to
communicate with. The two manners of accomplishing this are dial
pulse and touch-tone (a registered trademark of ATT by the way).
How efficiently touch tones are used depends on a variety of
factors.
In Electronic Switching Systems (ESS), the tones are received by
2 out of 7 tone decoders and fed directly to the processor, which
takes the necessary actions to place the call. If a user has a
dial phone, the pulses are counted by a scanning routine in the
ESS system and the same information (more or less) is used to
place the call. Clearly, tones are faster in this case.
In Crossbar exchanges (an older common-control switching system --
relay based via big XY matrix switches), the dial pulses are again
counted by specialized equipment (a relay bank in this case) and
fed to the call setup equipment. Touch-tones are decoded as in
the ESS system by 2 out of 7 tone decoders, and are fed directly
to this same equipment. The common-control, matrix nature of the
Crossbar office allows the direct use of this information. Once
again, tones win by far.
In Step by Step (SXS) offices, things are not so simple. These
offices (in their most primitive form) are based on "Strowger"
switches which pulse vertically and then ratchet horizontally
across a bank of contacts. These switches are inherently
designed for dial pulse operation, and were the first phone
switching technology. In a primitive SXS office, touch-tone
service can be provided by receiving the tones and converting
them to streams of dial pulses to drive the Strowger switches.
This can be done very simply by LSI circuits, complete with
digit buffering so that users need not worry about "dialing"
the tones too fast. In some offices, you can hear the pulsing
take place after each digit is "dialed".
However, there is a bizarre twist to this. While virtually
all Bell System SXS offices provide touch-tone service via
the mechanism mentioned above, many General Telephone offices
actually have a more advanced system. Instead of converting
tones immediately to dial pulses, they gather the tones directly
and hold them until the call destination is determined. If the
destination point does not require dial pulses (e.g. is not a
SXS office) the number information is transmitted via other
means (see (2) below) without any intermediate conversion to
dial pulses.
What this amounts to is that in a primitive SXS office, tones
gain you nothing in speed. In a hybrid office (such as many
General Telephone offices around Los Angeles) you actually gain
essentially as much as in a Crossbar or ESS office, since tones
are only converted to pulses when absolutely necessary.
There is also a switching system known as Panel, which is such
an abomination that I will not even discuss it. It is almost
completely non-existent in the U.S. today.
2) Point (1) above discusses getting the number information INTO
the switching system. Now we have the issue of getting the
information OUT to the destination.
If a call is local to an office, the information must simply be
translated to the correct directory/equipment number and the call
completed. In a Crossbar or ESS exchange, this is done directly
and the call is immediately completed. In a primitive step office,
the pulse dialing information is either used directly to complete
the call (in which case the call seems to complete instantly after
dialing the last digit), or possibly re-pulsed by the equipment
as necessary. In some cases, even out-of-the-office calls may
be dialed "directly" by pulse dialing in old SXS offices. In a
newer GenTel SXS office, the tone information is converted to
dial pulses at high speed to run the local Strowgers to complete
the connection.
If the call is going OUTSIDE the local office, the manner of
passing information depends on the DESTINATION office. Most
offices can accept incoming number information from other offices
in dial pulse form. Crossbar, ESS, and some SXS offices can take
it in MF form as well (high speed tone pairs not to be confused
with touch-tone). As time goes by, more and more offices will
be able to accept data via special digital data trunks that pass
number information for a whole group of incoming voice trunks.
If a call is going out from an ESS, Crossbar, or "smart" GenTel
SXS office, the information is sent via MF whenever possible,
since it is the fastest method. If the call is to a typical SXS
office, then dial pulses must be used, regardless of what sort
office is ORIGINATING the call. If the call is to a smart GenTel
SXS, MF may sometimes be used. Most toll (long distance) circuits
are MF based with an increasing number of them switching over
to the digital scheme mentioned above, which has a number of
advantages.
Overall, calls from ESS to ESS are faster, especially if they
are connected by digital address trunks. Calls between ESS and/or
Crossbar offices are also very fast. A standard SXS office at either
side slows things way down since dial pulses must be used. A smart
GenTel office can PLACE calls outside the office just about as fast
as a Crossbar or ESS office, but incoming calls will be slowed down
since the final connection must involve conversion to dial pulses
(though this may be done at 20 pulses per second instead of the
usual 10 per second).
Summary: If you live in a Crossbar, ESS, or smart GenTel SXS exchange,
Touch-Tone is faster overall. If you live in a primitive SXS office
(like an old Bell office, super rural GenTel or Continental Tel.
office), tones win you no speed, though possibly some convenience
in "dialing".
--Lauren--
P.S. A couple more points to clean up the loose ends. General
Telephone calls their computer based switching systems EAX
exchanges (for Electronic Automatic eXchange). They are pretty
much functionally equivalent to Bell ESS exchanges. In fact,
a rather surprising fact is that some Bell offices have bought
EAX equipment from GenTel and some GenTel offices use Bell ESS
equipment! MF stands for "multifrequency", in case you wondered.