What's new
iamretro

Wow! It seems you have yet to create an account with us! Needless to say, you have to register in order to get the full "iamretro" experience! You will then be able to create your own topics, reply to posts and be an active member of the MOST prolific community in Greece (even further than that, but we're being modest for now). So don't delay! REGISTER!

2001: A Game Odyssey

Leon

Retro Enthusiast
Comrade
Joined
Nov 10, 2019
Messages
395
Συνέχεια από το προηγούμενο post

The computer could generate three views of the environment which were displayed on three television screens and in turn were projected to three wide-angle window simulators through the Farrand optical system. The views created by the computer consisted of an infinite ground plane which was textured with simple repeating geometric patterns (representing for example the moon’s craters) and of a star field with a cyclic pattern of “constellations”. Additionally there was a distinct area of texture and color on the ground plane which represented the landing point. Alternatively, in lieu of the ground plane the computer could display a bounded plane surface, with similar textures, called the rendezvous point. The shape, size and color of the textured patterns could be selected by the user. The surface was transformed in real time each 1/30 of a second, matching the refresh rate of the NTSC standard, to provide a perspective correct view to the user looking at it from the spacecraft’s window.
NASA-1.jpgimage004.jpgGroundPlane02.jpg
Views of the textured surface. The red square is the "landing point"

The computer calculated digitally the scaling, the pitch and the yaw but not the roll. The calculation of all three axes was way too taxing for the hardware of the time to be done in real time, even if it cost millions of dollars. It was assumed the surface was parallel to the television’s scanlines. So instead Schumacker came up with a method of rotating the scanlines themselves (which in a normal TV are always horizontal) using the electromagnetic deflectors of the cathode ray tube. A very cumbersome and difficult technique but still cheaper. They also implemented an analog technique to gradually fade out the detail of the texture pattern as it gets further from the observer, and thus smaller and smaller, to avoid moire patterns.

All this impressive (for the time) visual magic was computed by a special purpose display computer which consisted of eight units and several auxiliary ones. The computer received data from the NASA computers, which were responsible for the vehicle simulation, and converted them to data to drive the visual display. The system was built with pre-TTL equipment, using Computer Control Corporation (3C) circuitry cards containing discrete components. The Program Control Unit (PCU) employed magnetic core memory of 512 48-bit words with 5 ms access. With a mind boggling speed of 5 mhz it executed the frame by frame (1/30 of a second) computation with the aid of 5 more units which were responsible for rotations, perspective calculations, screen filling, etc. The Maps and Video Processors Unit (MVPU) stored the texture data and determined the color at each point.
image010.jpgimage005.jpgimage007.jpg
Houston The NASA-1 computer at Houston

Now it seems we have color pixels and texture mapping back in 1964, while it is widely known that texture mapping was pioneered by Ed Catmull ten years later. Well...almost. Each scanline was divided in 256 raster elements, that does sound very much like a “picture element”, but not exactly. The ground plane was textured by a repeating pattern of tiles. The tiles are squares which are created by 64 smaller subtiles in an 8x8 array. These were not pixels, but analog small geometric squares. The NASA-1 computer calculated a perspective view of the tiled plane one raster segment at a time for each scanline. It computed what tile and which subtile the astronaut would see at each point looking into the ground plane and it sent the stored color to the display. The color of each tile wasn’t stored digitally (so they were not pixels), but was set by a potentiometer. And while it was called “texture mapping” in the final report of the simulator, it wasn’t a fully digital process, so Catmull’s texture mapping still holds.

Just before the product was delivered to NASA the development team indicated that it was possible to generate some simple objects to place on the ground plane, so the contract was extended to add this function. Using the NASA-1 computer and a special purpose device built by discrete components (integrated circuits were not yet available) they created a demo of a solid red triangle rendered in real time, that rotated and translated convincingly in three dimensions. The demo was shown in 1965, the first time we had true color pixels and solid 3D polygons. It was one small step for the NASA-1 simulator, one giant leap for computer graphics (sorry, couldn’t resist).
GroundPlaneminimalobjects.jpg
The ground plane (with LPD overlay) with minimal polygonal objects, representing the landing spot.

While the demo showed that with current technology it was prohibitively expensive to create real time 3d polygonal graphics of greater complexity, functional integrated circuits were just being completed by Motorola. They would be used in the upgrade of the Electronic Scene Generator in ‘67, adding polygonal capability. And although the Interim Visual Space-Flight Simulator was not designed for crew training it would come to be used for that as well. But this is a story for another time.

Sources:
John Lott Brown et al, Visual Elements in Flight Simulation, National Academy of Sciences, Washington DC, 1975
Alvy Ray Smith, A biography of the Pixel, Cambridge, Massachusetts, The MIT Press, 2021, Chapter 7
GE–NASA, Final Report, Visual Three-View Space-Flight Simulator, Contract NAS9-1375, GE, Defense Electronics Division, Electronics Laboratory, Ithaca, NY, Aug. 1, 1964
GE–NASA, Sixth Quarterly Report, Modifications to Interim Visual Spaceflight Simulator, Contract NAS 9–3916, GE, Defense Electronics Division, Electronics Laboratory, Ithaca, NY, July 1966
GE–NASA, Ninth Quarterly Report, Modifications to Interim Visual Spaceflight Simulator, Con-
tract NAS 9–3916, GE, Defense Electronics Division, Electronics Laboratory, Ithaca, NY, Apr. 20, 1967
Major A. Johnson, “Electronics Laboratory (1944 to 1993),” Progress in Defense and Space: A History of the Aerospace Group of the General Electric Company (M. A. Johnson, 1993)
James A. Lawrence, H. Edward Smith, The Role of JSC Engineering Simulation in the Apollo Program, Simulation Magazine, July 1991
David C. Christianson, History of Visual Systems in the Systems Engineering Simulator, NASA Conference Publication 3043, Ocean City Maryland, 1988
Matthew Baum et al, Virtual Image Out-of-the-Window Display System Study, Volume I and II, Farrand Optical Company, New York, 1966
John McLeod, Simulation at the Manned Spacecraft Center, Simulation Magazine, August 1966
Robert A. Schumacker, Brigitta Brand, Maurice G. Gilliland, and Werner H. Sharp, Study for
Applying Computer-Generated Images to Visual Simulation, report AFHRL-TR-69–14, Daytona Beach, FL: General Electric Company, Sept. 1969
C. H. Woodling et al., Apollo Experience Report – Simulation of Manned Space Flight for Crew Training, NASA TN D-7112, Washington D.C., March 1973
 
Last edited:

Leon

Retro Enthusiast
Comrade
Joined
Nov 10, 2019
Messages
395
Θα δούμε τον NASA-2 simulator κάποια άλλη στιγμή, μιας και είναι τεράστιας ιστορικής σημασίας αφού είναι ο πρώτος που χρησιμοποιεί shaded solid polygons. Στα 2 εκατομμύρια ήταν ένα λίγο ακριβό παιχνιδάκι αλλά 20 χρόνια μετά είχαμε πιο ρεαλιστικούς simulators στα PC μας. Πάμε στο Cambridge to 1968 όπου θα δούμε ένα άλλο παιχνίδι με μεγάλη, όπως θα διαπιστώσετε, ιστορική αξία και δεν του φαίνεται καθόλου εκ πρώτης όψεως.

357. MOO (1968) Developer: Frank H. King Platform: Cambridge Titan and others

The first video game with a high score table. One of the first cases of security hacking. A very famous 1970s board game. A modern day web game sensation. Tron. One of the most played mobile games of all time and the founding of one of the largest video game publishers. How are all these connected? Read on to find out.

I am sure you all know Wordle, which became a worldwide phenomenon in late 2021. Its creator, Josh Wardle, sold the game for an undisclosed seven-figure sum to the New York Times in early 2022 making him rich overnight. And he sold it just in time, before the fad (like any fad) died. Since then there was a steep decline in its popularity and soon it will be all but forgotten. So in case you read this in the far future (like 2023), Worlde is a code-breaker web game. Each day the players have six guesses to find a five letter word. After each attempt the computer tells them which letters are found in the word and if they are in the wrong or right place. Very simple but also an interesting problem for linguists and mathematicians. What made the game such a success was that you could share your result on social media comparing it with your friends. The daily competition plus the social aspect of the ensuing discussion (with what word did you start? In how many attempts? Etc.) made it extremely addictive.

Old-timers (like you who reads this) would compare Wordle to another fad of the ‘70s, Mastermind: “It’s like Mastermind, but with words”. Mastermind was a board game created in 1970 by Mordecai Meirowitz, an Israeli postmaster. The game itself was “like Bulls and Cows but with colored pegs instead of numbers.” Bulls and Cows is a pen and paper game of unknown origin, probably devised in the 19th century or even earlier. The code-master sets a 4 number sequence (like 9563) and the code-breaker must divine it with as little guesses as possible. A correct number in the correct place is awarded a bull and a correct number in the wrong place a cow. Meirowitz tried to sell the game to various game companies but was promptly rejected, since “who would pay for something they could play for free?”.

Turns out more than fifty five million people. The game was picked up by a small Leicester UK company, Invicta Plastics, when they saw it in the February 1971 Nuremberg International Toy Fair. Invicta at the time wanted to expand to the lucrative business of toys and games. As luck would have it one of its employees had played a similar computer game in Cambridge (Moo) and recognized the game’s potential something that the other companies failed to. They renamed it Mastermind, the same name as a trivia BBC show, and released it 1972. It caught the “adult party game” wave, like Scrabble and Monopoly and became an instant sensation. Instead of going to a fancy restaurant and paying through the nose, many groups of friends chose to host dinner parties and play board games (at least the conservative groups who weren’t into the normal ‘70s drug-fueled sex orgies). There were national championships (held in the Playboy Mansion in the States) world championships, it was used by the Australian Army as a training tool, and was owned by 80% of households in Denmark. In the ‘80s the fad died down and the game was retired to some dusty closet.
mastermind1.jpg
Mastermind's iconic cover targeted an adult audience

Before Bulls and Cows became a huge board game in the form of Mastermind, it had become a successful computer game in 1968 in Cambridge. It was written by Frank King for the Ferranti Atlas 2 computer. In 1961 Cambridge required a computer with a limited budget. Peter Hall, Ferranti’s Computer Manager, suggested to provide the Atlas computer processing units at work cost and then jointly develop a simpler and cheaper memory system and peripherals. The end result would be the prototype of a cheaper version of the Atlas computer for the market. It was named officially the Titan (Ferranti used names based on classical mythology), but then the marketing department renamed it Atlas 2. In Cambridge the original prototype name Titan stuck. In 1965 Maurice Wilkes, a member of the Cambridge team, discovered time sharing in MIT, and demonstrated it to the rest of the group using a teleprinter connected with a telex line across the Atlantic. So they decided to add a time-sharing facility which allowed the connection of interactive terminals. When the system was operational in 1966, a multi-user environment was available, one of the first of its kind.
camatlas2.jpg
Construction of the Cambridge Atlas 2 Peripheral and Tape Co-ordinators

Frank King at the time was studying for a PHD in Electrical Engineering. He wanted to implement something on the computer and thought a game would be desirable to practice his skills in machine code. Tic-Tac-Toe was too simplistic so he went for a puzzle game from his childhood. The other students told him he wrote Bulls and Cows but he called it MOO. The big innovation of the game was that there was an online league table (the first ever probably), which took advantage of the system’s multi-user capability, where player score was recorded. The competition, trying to beat your fellow students’ score is what made this game successful. Ten years later this feature would play a very important role in the rise of arcade video games. Having your initials at the top of the high score table at your local arcade, gave you the incentive to come back to the game again and again.

Συνεχίζεται στο επόμενο post.
 

Leon

Retro Enthusiast
Comrade
Joined
Nov 10, 2019
Messages
395
Συνέχεια από το προηγούμενο

The league table should be tamper-proof and at first it was thought it was. Then a user named GREMLIN appeared at the top of the table with ridiculously high scores. Someone had managed to crack the system’s security, which was a matter that went well beyond having bragging rights to a computer game. The hacker could just as easily change the game’s source code and worse tamper with the operating system itself and any data found on it. It was pointed it to him by the university’s informal computer security team that in the future someone could hack a bank’s security and make an electronic withdrawal. At first it was a friendly competition. King would fix the code and GREMLIN would crack it again. In the end King gave up because he had to focus on his PHD thesis. Members of Archimedes (the Cambridge mathematical society) were suspected and in the end the computer group managed to secure the system but only by hiring the culprit. He refused to explain how he did it until he was threatened with a permanent ban from the computer. He ended up working in Cambridge for the next fifty years. Turns out that IBM computers had similar vulnerabilities. So when Cambridge bought an IBM mainframe they implemented the MOO fixes to improve its operating system’s security.

The next step for beating MOO, was to write automated software which played the game for the user, thus the first game bots were created. Another level of sophistication was required: one private command to call another. A successful program was written by Dr. John Larmouth which used information theory and gave a good average of 5,24 tries close to the best human average which was 5,2. Then Barry Landy (who was on the team which designed the Titan computer) managed to stay at the top of the league table for a very long time with a program which either played the game automatically or helped him by analyzing the current game he was playing. But neither strategy was optimal. Around 1978-1979 John Francis made a complete analysis of the MOO game tree using two months of CPU time on a DECSYSTEM-2050 and finally arrived at an optimal strategy.

In 1970 John Larmouth (of the bot writing fame) visited Jerrold Grochow at MIT and convinced him to write a MOO game, which was popular in Cambridge, for MULTICS to show off the system’s capabilities. The game was an instant success with more than a thousand games played in two months by 75 people. In six months 3000 games were played by 150 people out of 500 who had access to the system. Eventually its popularity fallen off with the introduction of new games to the system, like Conway’s Life. Exactly like Cambridge many users tampered with the leader board, showing that Multics was unable to have a tamper-proof leader board unless the game became part of the system, which was against policy. There were security measures taken for such attacks, but they eventually died down on their own as the game’s novelty wore off and players became bored and moved on to the next thing. The story of MOO on mainframes looks very similar to Wordle. It became a fad in the online community, mathematicians jumped on it to solve it and eventually it faded away.

And so MOO jumped to the other side of the pond where it was discovered by Lane Hauck. Lane Hauck studied physics and engineering at UCLA and Cal State. He worked for Lockheed in the early ‘70s when he discovered minicomputers and decided they were the future. He tried to convince Lockheed to buy one, but they denied. So he spent 5.500$ of his own money to buy a Digital Equipment PDP-8, one of the few individuals to have their own private mini computer. He connected to it with a KSR-33 teleprinter by Teletype. He taught himself assembly and started playing the games which came along with the computer. A game which fascinated him was MOO. His wanted to share the game with his friends, so he turned to his other hobby, making complex technology simpler and affordable. He soldered 25 logic circuits and created a handheld electronic game which could play MOO. In ‘77 he licensed it to Milton Bradley which was released as COMP IV. His tinkering with electronic games landed him a job in Gremlin, where he would create Blockade, the first lightcycle game (as they came to be called from the movie TRON) and Hustle, which led to the evolution of the Snake game, one of the most played mobile games, with more than a billion players. We will see both games in detail later.
Comp_IV_by_Milton_Bradley,_Made_in_USA,_LED,_9_Volt,_Model_4751,_1977._(LED_Handheld_Electroni...jpg
Lane Hauck's Moo handheld game published by Milton Bradley as COMP IV

How is this connected to the founding of a major video game publisher? Lane left Lockheed and went to San Diego to work for Spectral Dynamics, an instrument manufacturer. And guess who his coworker was: Trip Hawkin’s dad. In the early ‘70s Trip was a high school student and went to Lane’s house to play with his PDP-8. There and then decided that computer games were the future and decided to start a computer games company by the time they were cheap enough to enter households. From then on he planned his entire career around that goal and in 1982 he founded Electronic Arts. Another great story which is also covered in detail.

As for the game’s creator, Frank King, he got his PHD and had a long academic career in Cambridge. Dr. F.H. King became obsessed with time and clocks from ancient sundials to atomic clocks. He was the designer of some important sundial clocks. He is Keeper of the Clock, responsible for Cambridge University's official timepiece on Great St Mary's Church and chair of the British Sundial Society.
_105548479_clock.jpg
Dr. Frank H. King with Cambridge's clock mechanism
And that’s it for MOO, a simple but involving (especially for mathematicians) puzzle game with a very rich history. It was the first game to feature an online league table, the first game where someone used a cheat and the first game where bots were used. The cheat is on of the first hacks (as we mean them today) in history. The cat and mouse game of cyber security vs exploiting vulnerabilities started in Cambridge and later MIT over a game’s high score. The stakes today are much, much higher. You can find some variation of MOO on pretty much every computer under the sun and even most early consoles had a version of the game. The Atari VCS version was called Codebreaker and developed by Larry Kaplan. It made use of the Atari Keypad Controllers.


 

Leon

Retro Enthusiast
Comrade
Joined
Nov 10, 2019
Messages
395
358. Solar Fox (1981 Arcade, 1983 C64) Developer (C64): Judy Braddick, Cyndie Merten Publisher (C64): Commodore Electronics Ltd. Platform: Arcade, Atari VCS, Commodore 64

You might know Solar Fox as an Atari VCS game and be unaware of the fact it was first an arcade game. And why would you? It was a flop of monumental proportions. The copyright on-screen suggests that it was created in 1981, but it was most probably released in 1982. Arcade flyers, magazine references and the operator manual all seem to agree. Even worse, Bally might have shot themselves in the foot and released the game in the same month with their far more successful Tron, effectively killing the game. The game is like a cross of a maze game like Pac-Man where you collect stuff to clear the screen and a shmup. You control a spaceship moving in a grid and must collect items (Fuzors) while enemy sentinels (Thrustors) moving at the edges of the screen fire at you. When you collect everything you move to the next level. You have two buttons: a fire button and one which controls your speed. This maze shooter was considered too simplistic, too similar to previous games and too easy; a player could learn the patterns and play for a long time with one quarter, something which definitely didn’t appeal to arcade operators. It’s a shame really, because Solar Fox is a very enjoyable arcade game.

The simplified Atari port (you couldn’t shoot back) fared much better than the arcade original and it’s considered one of the better games of the console. CBS Electronics, the short lived video game branch of the giant conglomerate, planned on releasing the game on many other platforms but the video game crash happened and the company pulled out of video games. What is less known is that another port of the game exists for the Commodore 64, developed and published by Commodore. The company had an excellent relationship with the goodfellas of Bally Midway (a story for another time) and licensed many of their coin-ops for the VIC-20 and the Commodore 64. The C64 port of the game is a pretty good one and was programmed by a female engineer, something quite rare those days, Judy Braddick.
Casual '80s sexism in the Atari VCS commercial. Ironically, the C64 was done by two women.

Judy Braddick after graduating from college with a degree in computer science got hired in Commodore’s Recreational department, meaning games. She was surprised they hired her since Judy wasn’t a gamer. She had to pick up the required skills and fast. If one must write games one must know how to play them after all. They had a room full of Bally Midway arcade games, which were to be converted to the Commodore computers. She spent many hours there doing “research” and realized she had a knack for it, becoming quite good at Solar Fox and Satan’s Hollow. She became addicted to gaming; her friends had to drag her out of arcades where she “checked the competition”.
Braddick.jpg
Judy Braddick at the 1985 CES

She started as a trainee working for a senior programmer, in this case Andy Finkel. Her first real assignment was to write a routine for the game Andy was working on, DragonsDen, a procedure called black boxing. A black box is when a senior programmer tells to the junior what the routine must do, what data goes in and what must come out. In that case it was a flock of birds circling and attacking the player.

One day Judy was in the arcade room playing Solar Fox when John Mathias, the head of the group, asked her if she liked the game and if she wanted to convert it for the C64. She liked the game alright but she didn’t feel confident yet to take on such a project. Mathias reassured her he wouldn’t throw her at the deep end of the pool if she wasn’t ready. A month passed where she was writing subroutines for other programmers and playing Solar Fox more and more, but no news about the game. Then Mathias declared that either Judy or another coder, Cyndie Merten, who had just finished writing a demo cartridge, would port the game. It was decided amicably between the two with a coin flip.

As per usual the coin-op manufacturer didn’t provide anything more than the arcade machine and the service manual. No design documents or source code. So Judy had to play the game to death and faithfully recreate it. The first problem she encountered was screen orientation, since the arcade had a vertical one. She had to redraw the background six times before she came to a compromise which fit the horizontal TV screen but was also somewhat faithful to the arcade. The sentinels at the top and bottom of the screen had to overlap with the score and the text in order to make the grid taller. For controls she wrote three routines until she was happy with how it played using the Commodore joystick. So in case you still have one lying around that’s the ideal joystick for Solar Fox.

The biggest problem she faced was which objects would be sprites and which would be characters. When sprites move they don’t erase the background, but characters do. The C64 has 8 sprites available at any one time but Solar Fox has 16 moving objects. She made the spaceship and its missile a sprite, the four sentinels (Thrustors) at the edges, the blue gear (the Pulzor) and the little guy, the Freezor (which when shot freezes all the sentinels). That left the 8 vortices, the missiles the sentinels throw at you, be drawn by characters. That was a problem since as they moved across the screen they erased the grid. If the character had the grid graphics included (to create the illusion it was not erased) it surpassed the 256 character limit. If she left the grid out completely it looked bad. If she left the grid out and the had it drawn on top of the screen it made the game too slow. Two months of work down the drain.

The crisis was averted with the help of her supervisor, Andy Finkel. The solution was to multi-sprite or as an Atari VCS coder would call it, to race the beam. Basically you have to know the timing of the electron gun as it hits the TV. You draw the sprites at the top of the screen, erase them and redraw them at the bottom before the raster beam reaches that spot, creating the illusion of having more sprites. David Berezwoski had used that technique to draw as many as 64 sprites in a demo and in his game Satan’s Hollow had 32. So she changed the sentinels to character graphics, the Pulzor and the Freezor shared one sprite, the spaceship and the missile another two and she multi-sprited the vortices.

The last problem she faced was that no matter how hard she tried she couldn’t reach the final four levels to see the screen pattern. They had installed a freeze switch to pause the game in order to copy the each level’s graphics at her leisure, but it didn’t do any good if she couldn’t reach that level. The high-score screen says “designed by Arcade Engineering” a company founded by Ron Haliburton and Jack Pearson (Solar Fox’s designer), which had also developed other games for Midway, like Omega Race and Two Tigers. The company was later sold to Midway. So she called the engineers at Arcade Engineering and explained her predicament. They sent her a ROM which disabled the baddies, so she could go through the entire game with ease. The guy who helped her was Tim Gilbert, an Arcade Engineering programmer who did Two Tigers.

With all that work and the deadline approaching Judy didn’t have to do the sound. Commodore wanted this game on time. The sound effects were programmed by the coder who lost the coin flip, Cyndie Merten, making Solar Fox for the Commodore 64 probably the first game to be programmed by a team of female only engineers. The game was delivered with only a small two week delay, something of a first for the group.


Braddock, Juddy, The Making of a Game Programmer, Commodore Power Play magazine, p. 70, 03/85
Stone, Andrea, The House that Pac built, Play Meter, p. 53. Vol. 1, n.3, 12/82
Company Profile, Play Meter, p. 66, Volume 20, 08/94
Pearson, Jack, Universal Video Game Platforms, Play Meter, p. 48, Volume 21, 12/95
Solar Fox, Parts and Operations Manual, Bally/Midway, Illinois, September 1982
Mayers, Richard, Arcadia, Video Valhalla, Video and Computer Gaming Illustrated, p. 34, Issue 12, Dec. ‘83
https://hackaday.com/2015/01/10/making-the-ces-show-thirty-years-ago/
 
Top