Pleasant@Rutgers.arpa (02/07/83)
HUMAN-NETS Digest Monday, 7 Feb 1983 Volume 6 : Issue 7
Today's Topics:
Technology - EFT (2 msgs) & WorldNet (3 msgs),
Programming - Unix (5 msgs),
Computers and People - Human Memory Capacity
----------------------------------------------------------------------
Date: 5 January 1983 06:39-EST
From: Robert Elton Maas <REM @ MIT-MC>
Subject: EFT etc.
Tonight I got robbed at knifepoint. They took my wallet with cash
and credit cards after threatening me with the knife and then
punching me in the face giving me a big bloody nose. I yearn for the
days when we'll not have to carry cash or credit cards, when
finances will be done electronically by password, so there'll be no
incentive for street robbery.
------------------------------
Date: 6 January 1983 23:24-EST
From: Robert Elton Maas <REM @ MIT-MC>
Subject: along the lines of EFT??
Sigh, I got my broken nose fixed up (just a little chip broken loose
near the top) then renewed my driver's license to replace the one in
the stolen wallet. I then went to Macy's to try to buy a new wallet
since it was a pain to keep dollar bills loose in a pocket and coins
in a little paper envelop in another. But with my automated teller
card stolen I couldn't draw out cash, and Macy's wouldn't let me pay
by check (need two forms of ID, one with picture and other a major
credit card) nor by Macy's credit card (by the time initial
footsying around with them was done it was after 5pm so they
couldn't even set me up with a temporary credit card), so I'm still
without a wallet.
I don't like the current system where one must carry around papers
to prove you are a valid person and if you get robbed you are a
non-person for several weeks until your driver's license (6 weeks)
and credit cards (2-3 weeks) are re-issued. Why can't they accept
thumbprints as an alternate way of identifying people? Or why can't
they connect with TRW or other major credit firm and identify I'm
really me by asking me personal questions that aren't known to
anybody except me and the credit agencies? I hope things get better
not worse in the future. Of course like I said before, not having to
carry around cash would reduce the incentive to get robbed in the
first place, and as I'm adding now, if everybody carried around a
little radio transmitter that detected your pushing a button
(explicit emergency) or falling to the ground or losing vital
functions (implicit emergency) and sent out a call for help, maybe
we'd be able to deter crime by catching all muggers immediately when
they try anything.
Anybody want to brainstorm about how such things might be done with
packet radio or whatever, using current technology plus the effort to
create an actual network in various high-crime areas and
high-population cities? Anybody want to warn about some of the
Orwell-style misuses possible with what I claim I want?
------------------------------
Date: 27 January 1983 04:50 EST
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: Changing face of u-Computing.
Well said. My point precisely, stated somewhat better than I put
it. My father owned radio stations most of his life, but he never
knew anything about electronics. I can change sparkplugs without
knowing much about the theory of Kettering ignition and how the
distributor works.
The Epson QX-10 is a case in point. The software isn't
implemented yet, but when it is, damned near ANYONE will be able to
use the machine to do a LOT..
------------------------------
Date: 31 January 1983 02:58 EST
From: Robert Elton Maas <REM @ MIT-MC>
Subject: Worldnet & Lawyers
I'm afraid I have to disagree and say that keyboard entry and CRT
reading is more like written word than like spoken word because it
uses representations of written letters an spelling conventions
instead of representations of spoken phonetics. Of course in
languages that have phonetic written languages this distinction
breaks down, but if we use English and other non-phonetic languages
as testbed we indeed find they all use written language rather than
phonetics. (Correct me if I'm wrong.)
------------------------------
Date: 6 Feb 83 17:40:35 EST (Sun)
From: Chris Torek <chris.umcp-cs@UDel-Relay>
Subject: Re: Worldnet & Lawyers
From: Robert Elton Maas <REM@MIT-MC>
I'm afraid I have to disagree and say that keyboard entry
and CRT reading is more like written word than like spoken
word because it uses representations of written letters an
spelling conventions instead of representations of spoken
phonetics. Of course in languages that have phonetic written
languages this distinction breaks down, but if we use
English and other non-phonetic languages as testbed we
indeed find they all use written language rather than
phonetics. (Correct me if I'm wrong.)
I think that's irrelevant. The trouble is that keyboard entry is
more like written things since it is easily reproduced. Once you've
said something, then (unless someone has a tape recorder) it's gone.
If you've written it, then it's there on paper, evidence. This
makes (for example) the distinction between libel and slander.
Something you've said can be slanderous, whereas if you'd written it
it'd be libelous. (I could easily have those backwards. I'm not a
lawyer.) You might claim this is splitting hairs, but it's what
makes lawyers rich. Anyway, my point is that while stuff I type is
more like something written than something said, in that it can be
reproduced, it's less like it in that I'm much more CAREFUL about
what I write than what I type in. I'd rather not be charged for
libel against Foobar if I sent e-mail to the net saying "Foobar's
widgets are worthless."
------------------------------
Date: 24 Jan 1983 1136-MST
From: Walt <Haas@UTAH-20>
Subject: Re: TOPS-20 vs VMS vs UNIX
One OS environment that Tony <Li@RUTGERS> forgot to enumerate is
real-time process control. VMS appears to me to be superior to Unix
for this application. The ideal combination in this environment
would be, as Gail Zacharias pointed out, VMS with a Unix emulator to
keep the hackers happy.
------------------------------
Date: 24 Jan 83 09:13:38 PST (Mon)
From: sdcsvax!sdchema!donn at Ucb-C70
Subject: VMS emulation on UNIX
Speaker-to-Animal's point (Vol 6, #5) that no one would want to use
VMS facilities when they had UNIX is not quite enough to explain why
VMS has never been emulated under UNIX. There are in fact some
utilities that VMS has which users miss under UNIX. One of these is
VMS's highly optimizing FORTRAN compiler. Unfortunately DEC is much
more secretive about source code than Bell is, so much so that it is
close to impossible to buy machine-readable source for most VMS
programs. Hence the route which most UNIX emulations on VMS follow
fails for VMS emulations on UNIX: you can't get the source. Our
group at UCSD has coveted the FORTRAN compiler for some time and at
one point we proposed to DEC that we would write a program for UNIX
which would take VMS object modules and convert them into UNIX
modules which we could then load with a compatibility library and
run under UNIX. This would allow DEC to sell VMS FORTRAN and other
compilers to UNIX sites which they would otherwise fail to make any
money off of, and it would specifically let us dump VMS on the only
remaining machine here which runs it. DEC in its infinite wisdom
foresaw that they would make less money from sales to UNIX sites
than would cover the costs of many sites dumping VMS (like us), so
they have refused to give us objects or even symbol table layouts.
So much for VMS emulation.
Donn Seeley UCSD Chemistry Dept. RRCF ucbvax!sdcsvax!sdchema!donn
(619) 452-4016 sdamos!donn@nprdc
------------------------------
Date: 27 Jan 1983 1937-PST
Subject: VMS vs Unix
From: Mike Leavitt <LEAVITT at USC-ISI>
Why would I want to run VMS under Unix? Well, I would like it
because there is commercial applications software that I need that
is not (yet) available under Unix, and the systems types (pace JB)
who make such decisions for our agency's leaders have determined
that Unix running under VMS is "just too slow." Give me one or the
other -- that's not too much to ask!
------------------------------
Date: 29 Jan 1983 2306-PST
Subject: [Michael C. Greenspon <ZZZ.MCG at MIT-OZ at MIT-MC>: Re: RSTS
Subject: dies?]
From: Ian H. Merritt <MERRITT@USC-ISIB>
I received this interesting tidbit this evening, and thought it
relavent to the recent discussion on the subject of UNIX vs VMS vs
TOPS-20 vs ...
<>IHM<>
---------------
Return-path: @MIT-MC:ZZZ.MCG@MIT-OZ
Date: 29 Jan 1983 2321-EST
From: Michael C. Greenspon <ZZZ.MCG at MIT-OZ at MIT-MC>
Subject: Re: RSTS dies?
To: NCP.EGK at SU-GSB-HOW at STANFORD-GATEWAY
cc: gutfreund.umass-coins at UDEL-RELAY, info-rsts at MIT-MC,
merritt at USC-ISIB
Hey, watch it! I'll be the first to admit that RSTS is old,
generally ugly, and very stupid about lots of things, but it IS
hackable. UNIX is equally, if not more, ugly, stupid, etc., and is
also hackable. The difference is that in order to get UNIX to do
ANYTHING even REMOTELY USEFUL it MUST be hacked. Of course, if you
like case significance and an operating system designed around the
same mindless philosophy, go ahead and use stock UNIX. For now, on
PDP-11s, I'll stick to hacking RSTS.
Flame, flame, MCG
P.S. I suppose when the [rest of the] world gets color workstations,
someone will hack up a UNIX shell that is not only case significant,
but font significant, color significant, etc... Lotsa luck.
------------------------------
Date: 30 January 1983 16:57 EST
From: Robert Elton Maas <REM @ MIT-MC>
Subject: TOPS-20 vs VMS vs UNIX
Funny how the meaning of "hacker" changes from day to day according
to the needs of the person writing the message or article that
contains that word. By some standards I'm a hacker, yet I don't
spend lots of time modifying my user interface (HACTRN and init
files here on ITS). I wonder if there's any real concensus what the
word "hacker" means? Does it mean:
- Somebody who spends long hours modifying the user interface?
(Used this way in the message I'm replying to)
- Somebody who learns the innards of programming systems in
order to take advantage of optimizations and special methods
that aren't described in an documentation? (A favorite MIT
definition I've heard.)
- Somebody who likes computers and dislikes humans so much he
avoids all social contact? (Used this way in an article in a
trade journal last summer.)
- Somebody who writes programs by hook or crook to make them
work without necessarily organizing the programs in clean ways
and without conforming to any programming standards or using
standard algorithms for solving the tasks? (My favorite
definition.)
Meta comment -- A common way to lie is to have a word defined two
different ways. First we prove one definition fits the circumstance
at hand. Then we start using the word without reminding people
which definition we're using. Finally we start slipping to the
other definition, "proving" things based on that definition. I fear
this practice, already common in politics, may slip into our
discussions if we aren't careful.
So, I'd like to hear from you all, what you think the best
definition of "hacker" should be. Send your replies directly to me
and if I get enough of them I'll edit a summary and send it out to
the list.
------------------------------
Date: 30 January 1983 1649-EST (Sunday)
From: Alex.Rudnicky at CMU-CS-A
Subject: human memory
Some time ago, someone on this list stated that the ability to
memorize large random-dot patterns provides evidence that humans can
store large amounts of (visual) information. In fact, there is
considerable doubt that this is the case. The experiment in
question is reported by Stromeyer & Psotka [Nature,1970,v.225,
p.346]. The experiment involved the binocular fusion of Julesz
patterns ( two slightly different random-dot patterns are presented,
one to each eye, the disparity is such that viewers see a
(binocular) three-dimensional figure emerge from the pattern. Same
principle as 3D movies, only a meaningful image cannot be extracted
from just one pattern.) "Normal" viewers can integrate such
patterns presented in temporal sequence, but only if the patterns
follow one another within about 150 msec. The individual described
by Stromeyer&Psotka claimed to fuse patterns presented up to 3 days
apart (depending on the number of dots involved). The experiments
have never been replicated. I have also been told, by someone
acquainted with the individuals involved, that there is a good
chance that the young woman in question, by all accounts
exceptionally intelligent, may have been able to bluff her way
through the whole thing (in any case, shortly after the experiment,
she claimed to have "lost" her ability). Even if no deception was
involved, it is by no means clear that the amount of information
retained by the viewer is equal to the number of dots in a pattern.
In his own work, Julesz reports that fusion effects can be obtained
even when the patterns are substantially blurred (ie, every single
dot need not be remembered, low-frequency information is
sufficient.)
Incidentally, several people have described memory capacity in terms
of the number of (eg) pixels in a glance, etc. This is not the
correct way to describe how humans gather information. Think about
it. You do not perceive the world in terms of pixels. You are not
aware of the individual receptors on your retina, nor of the
impulses traveling along the optic nerve. Human processing is
highly selective, it does not retain the raw products of sensation.
You experience and remember objects, ie meaningful entities, but not
their meaningless (and much more numerous) constituents.
In a slightly different vein, I find that discussions of "how big is
human memory" are essentially beside the point. Mere capacity is
not what makes human memory interesting. How is it organized, how
is knowledge stored and retrieved, how does forgetting take place?
These are the questions that need to be answered.
------------------------------
End of HUMAN-NETS Digest
************************