textfiles/computers/CYBERSPACE/farmerge

Article 92 of sci.virtual-worlds:
Path: milton!randy@xanadu.com
From: randy@xanadu.com (Randy Farmer -- A survivor of the Lost Patrol)
Newsgroups: sci.virtual-worlds
Subject: Cyberspace: Getting There From Here **(LONG)**
Summary: Repost of article posted in alt.cyberspace, by request
Keywords: cyberspace hardware Habitat Club Caribe
Message-ID: <1990Mar8.160311.3103@xanadu.com>
Date: 8 Mar 90 16:03:11 GMT
Sender: hlab@milton.acs.washington.edu
Organization: Xanadu Operating Company, Palo Alto, CA
Lines: 289
Approved: hitl@hardy.u.washington.edu


I have had a few requests to repost my article, so heregoes!

I would really appreciate it if you sent me your thoughts on this text via
email (to save net bandwidth), and resist the urge to fragment my thoughts
in a followup posting with lots of '>'s. This article is meant to convey one
thought: Cyberspace is NOT hardware. Thanks :-)

This article originally appeared in the Journal of Computer Game Design last
October, and is positioned as kind of primer. Those of you who received a
copy of this at Hackers 5.0 or read it in alt.cyberspace, hit 'n' now :-).
----------------------------------------------------------------------------

               Cyberspace: Getting There From Here.
              (c) Copyright 1989 F. Randall Farmer

Randy was lead C64 programmer on the fabled 'Habitat'(1) project 
for Lucasfilm Games. He has taken a hiatus from building Cyberspace 
while he works on a new telecommunications product at AMIX, a 
subsidiary of Autodesk. He can be reached via Usenet as 
randy@xanadu.com or through the Journal BBS.

	"Golden Vaporware" is a term I've heard Ted Nelson(2) use to 
describe such exciting computer projects as Xanadu, Dynabook and 
Cyberspace. I understood the term to mean Really-Great-Stuff-That-
People-Are-(Sort-Of)-Working-On-And-Never-Ship.  But now, it 
seems,  "The Age Of Golden Vaporware" has arrived. Ted's Xanadu 
project is now well under way and should ship product in '90.  There 
is a company called Dynabook Technologies that has released "step 2" 
of the "20 steps"(3) to Alan Kay's vision of the ultimate laptop.  And 
Cyberspace has been prototyped by QuantumLink and Lucasfilm.

	These projects are likely to succeed not because they are 
exciting to us techie-geeks, but because they have solid commercial 
and technical foundations.  Here I will illustrate a path from current 
software and hardware technologies to the Cyberspace that many of 
us have dreamed of for almost 20 years.


What is Cyberspace anyway?

	"Cyberspace" is a fuzzily defined word that has been in the 
press a lot recently(4)(at least in Silicon Valley).  Vernor Vinge first 
described a vision of what might now be called Cyberspace in his 
novella True Names(5). The hero of Vinges's story connected to "The 
Other Plane" using EEG electrodes placed on the forehead. The Other 
Plane was a place where complicated software systems were 
represented by familiar objects.  The Max Headroom television show 
used similar representations in various episodes.  Other fictional 
treatments include William Gibson's Neuromancer (6), which has 
been turned into a popular graphic adventure, and other 
"Cyberpunk" novels. Here the vision of Cyberspace is based around 
commercial and governmental network computer access, where 
"cowboys" and "wizards" break in and wreak fiscal havoc. In these 
works, Cyberspace is a metaphor for abstract concepts.

	Autodesk, maker of the best selling AutoCAD system, has 
thrown its hat into the Cyberspace ring where it joins VPL's Jaron 
Lanier and NASA-AMES. All of these products share similar base 
technologies and assumptions: two head-mounted TV monitors for 
stereo imaging, a powerful computer, and a 3D input device like 
VPL's DataGlove. They are primarily single user systems, and, in the 
case of at least Autodesk's product, probably aimed at CAD users who 
want to edit drawings in three dimensions. There are many other 
applications, but discussion is beyond the scope of this article. The 
key to all of these systems is enough computer power to do  double-
buffered stereoscopic real-time graphics at a high frame rate (say 
20-30 frames/second). This technology is expensive: in the case of 
VPL's "virtual reality" system (their special two user version) the 
price tag weighs in at $69,925. The "Eyephones" cost $9,400 and the 
glove $8,800.(7) These companies envision Cyberspace as a new kind 
of user interface.

	To me, Cyberspace is a place, not just an interface or a 
metaphor. A place where people, regardless of location, hardware, or 
purpose can get together in a participatory experience to conduct 
business, socialize, or have a good game of SpaceCombat9.6.

	Why this vision instead of the others?  People.  Unlike the 
interactions with artificial personalities created for computer 
adventure games, Cyberspace interactions could create relationships 
that are "greater than the sum of their parts".  Here the consequences 
of individual actions take on a greater significance because they 
affect the world of, and the personal belongings of the other 
participants.  Because Cyberspace will be so malleable,  each 
individual can "make a dent"(8) and participate in its ongoing 
creation and adaptation.

	This is not speculation! During Habitat's beta test, several social 
institutions sprang up spontaneously: There were marriages and 
divorces, a church (complete with a real-world Greek Orthodox 
minister), a loose guild of thieves, an elected sheriff (to combat the 
thieves), a  newspaper (with a rather eccentric editor), and before 
long two lawyers hung up their shingle to sort out claims.  And this 
was with only about 150 people! My vision encompasses tens of 
thousands of simultaneous participants.


How can it work?
	There are several major problems facing a large-scale 
Cyberspace system: Bandwidth, Graphic Resolution, User Interface 
Standards, Event Integrity, Data Communications Standards, and 
Computer Horsepower.

	Bandwidth is the most overrated problem. Everybody thinks 
you need to send megabytes of data every minute to each user in 
order to make Cyberspace work. This is simply not true. Habitat 
works at a mere 300 baud. The keys to keeping bandwidth down are 
distributed processing (having the local computer do most of the 
display and interface work), object oriented command messages, and 
avoiding communication intensive graphics primitives (such as those 
associated with NAPLPS, that slower-than-molasses protocol they use 
for Prodigy and other Videotex systems).

	Graphic Resolution, User Interface and Local Horsepower issues 
should all be lumped together.  To be viable over time, Cyberspace 
must be designed with the "to each according to his abilities" 
principle. If a C64 with NTSC TV and joystick-only interface 
encounters a 60 megahertz 486 user with Eyephones and DataGlove, 
the interaction should be filtered so that each user can have a 
satisfying experience (e.g the C64 user doesn't see all the intricate 
detail the 486er has put into his attire, and all people walk exactly 
the same, but they can still sit, chat, and play a nice game of 
Strattagema.)

	Event Integrity is a side effect of distributed processing. Simply 
put: you can't trust the home computer. Hackers have all the time in 
the world to crack the software and make it send messages it 
shouldn't. This is akin to the long debated copy protection problem, 
but comes with its own unique solution: The Host. Since the host is 
the final arbiter of all events, it should be programmed at a very 
high level to reject bad messages.

	Data Communications Standards are key to the future viability 
of Cyberspace. A protocol is required that is efficient and forward 
looking,  with hooks for future revisions (as bandwith increases) and 
data encryption. Anything that is going to be used for business will 
need to transmit its data securely for protection against espionage.  
Also, the standard needs to address such issues as Email and  
gateways to those entrenched text-only systems.  I want to read my 
Usenet mail printed on a sheet of paper in my Cyberhand!

	Since the host coordinates all activity for all users, Host 
Horsepower is by far the most critical issue. For the first few years of 
operations, Cyberspace will have a single centralized host consisting 
of several connected multi-processing computers.  In order to keep 
overhead to a minimum, bandwidth and database access will need to 
be tightly limited.  The communications protocol will need to be 
super efficient and messages must be kept tiny.  Eventually there 
will be several independent Cyberspaces, perhaps on LANs or a 
larger BBS. Of course, while you are walking around in your office 
Cyberspace, you might want to go to other Cyberspaces, so a 
distributed host model will need to be designed.  By not having a 
central host keep track of everything and everyone, this distributed 
model will actually decrease the load on each host. This will allow 
Cyberspace to take on a few interesting features: 1) Each host can be 
a different kind of place, with different 'rules' (e.g. A role-playing 
Cyberhost would probably have monsters, combat rules -including 
character death- and a no-holds-barred policy about participant 
behavior whereas an office Cyberhost would have access protection 
and wouldn't allow personal combat of any type), 2) Increased 
capacity without loss of access, and 3) Faster growth.

When will it happen?

	Rather than pin a date on my Cyberverse, I will outline a 
possible implementation path, including all of the developments 
publicly announced to date.

Event: Spring 1989 "Habitat" A.K.A. "The Poor Man's Cyberspace"
	Ship Low End Prototype:
		Third person, 2D, Low resolution, Low bandwidth,
		multi-user Cyberspace with joystick/keyboard interface
		Purpose: Entertainment

Event: Fall? 1990 "CyberCad"
	Ship commercial CAD version of Cyberspace Interface
		Eyephones, DataGlove/DataSuit. First person, single-user
		3D graphics (wire frame on small computers).
		Purpose: Commercial CAD
		
Event: AT&T starts installing nationwide ISDN(9), promises 
installation
	by 2001 (this should drive bandwidth to 1.5gigabits/second!)
Event: 'DataGlove, the cheap version' arrives.
Event: 9600/19.2k baud modems <$100 dollars
Event: A Cyberspace team defines the first Cyberspace Data and 
	Communications Standards (an event they will regret later
	when this interim hack becomes the defacto standard)
Event: 486/68040 50+ megahertz computers sell for <$1000 dollars.

Event: (about 1995) "Cyberverse 1.0" (U.S.? Japan?)
	Ship first multi-purpose Cyberspace universe
		Eyephones optional, mouse or glove required
		Local data stored on high density media (CDROM?)
		9600baud minimum speed, LAN version available.
		(host still required)

Event: Computers get 10X faster and 10X more memory/storage.
Ongoing event: AT&T still installing ISDN
Event: (about 2000) Japan completes ISDN installation, selling
	ISDN 'modems' cheap!
Event: First 'cracker' group has successfully done measurable
	damage to the global Cyberverse.
	The back doors are all closed. Or are they?
Event: After cursing the original Cyberdesigners for several years
	now, the Data and Comm standards are updated to support.
	multi- and distributed-host models. Massive testing required.
Event: First suicide attributed, by the media, to a game played while
	in Cyberspace.
Event: First combination television/phone/computer successfully 
	mass marketed??

Event: (about 2000) "Cyberverse 2.0"
	Ship distributed host version
		Supports latest user interface hardware (optionally).
		Now a host is not required.

Event: (about 2010) AT&T finally completes ISDN. Wall size 
	videophone reaching suburban markets (10-20 years late?)
Event: Congress takes significant note of the Cyberverse, because
	trade unions (and other lobbies) note significant
	changes in the distribution of political power.
	First Cyberspace lobby arrives in Washington.
Event:10-20% of Americans spend over 4 hours per day "on the
	Other Side". First chapter of Mothers Against Cyberspace
	forms, claims 'Junior' never goes out to playI
	(We gotta think about this stuff too, ya know!)

Event: (the Future) "Cyberverse 3.0" A.K.A. "The Big Payoff"
	Distribute version 3.0 via Cybersoft retail outlets in Cyberspace
		100,000+baud, allowing life-like personalizations
		EEG direct input? Hologram Video? Retinal Projection?
		Full Body digitizing? Forced feedback?
		Voice input/output? World access via remote-robots?
		

Farmer, you're full of Cybercrud!(10)

	Oh, yeah? A 386/CDROM version of Habitat will soon be 
released in Japan.  Other foreign companies are also interested in this 
technology, and some are talking about multi-machine support, 
standards, and the future.  My only fear is that the long-term nature 
of this project will cause American companies to shy away, and let 
yet another new technology be monopolized by other, more forward-
looking nations.

(1)Credit where credit is due: Chip Morningstar was original designer 
and leader of the Habitat project. Aric Wilmunder was the other 
primary C64 programmer. Janet Hunter did most of the Host work. I 
was also chief operator and world builder. Habitat has not yet been 
released in its complete form (it will in Japan in '90). For a taste of a 
much-scaled-down version, check out Club Caribe on QuantumLink 
(you'll need a C64 & software). Call 800-782-2278 for details.
(2) Nelson, Theodor, Standard Public Xanadu Speech. Xanadu is the 
name of the original mega-hypertext product.
(3) Khosla, Vinod, of Dynabook Technologies quoted in Bay Area 
Computer Currents "Dynabook! Dynamite?" Vol 7,#4 pp 22-24.
(4) Laurel, Brenda, mentions Cyberspace in "New Interfaces for 
Interactive Entertainment" in Vol 2, #5 of The Journal of Computer 
Game Design.
(5) Vinge, Vernor, "True Names" republished in an anthology "True 
Names and Other Dangers" (mandatory reading for would-be 
Cyberhacks)
(6) Gibson, William, "Neuromancer"  (recommended reading)
7)Nix, Shan, interviews Jaron Lanier of VPL in The San Fransisco 
Chronicle, "Welcome to the World of Hyper-Reality" early August 
1986, pp B1,B6
(8) ibid
(9)ISDN:Integrated Services Digital Network: instead of converting 
digital data to analog (via modem) and back again, digital data is 
dropped directly on to phone lines. Maximum throughput of full scale 
fiber-optic ISDN is 1 to 6 gigabits/second or more. Several foreign 
countries have made ISDN a national priority.
(10)Nelson, Theodor: Computer Lib/Dream Machines Microsoft Press 
Revised Edition 1987 pp 27-29

----------------------------------------------------------------------------
Well, there you have it, my first published article :-). All comments welcome.

Don't pay too much attention to the timeline, it's meant mostly to point
out that there are no real technical barriers to this vision, and the future
should only see the kind of improvments we've seen with computers: Faster,
Bigger, and more Resolution.