Lost From 1983: Garfield’s Quest by Atari

arcadehero April 2, 2020 0
Lost From 1983: Garfield’s Quest by Atari

Now that April Fools is past us (until next year 😉 ), we can get back to real news. Although this post is dealing with a new find of an old game that never came into being. I love finding information on long lost prototypes and sharing that data here, although it has been a while since we had a new discovery in the arcade space.

Today’s game once again comes from the once-prolific game designers over at Atari. When you look at their list of almost finished games, you think that everything has already been discovered, then something like this pops up out of the blue. The discovery was shared on the AtariAge forums by user Dutchman 2000, who in looking at his posting history, spends a lot of time searching for game treasures and trying to preserve them. He states that he found the data for this title on a 9-track data tape; unfortunately no artwork, screenshots or video appear to have been preserved, although it is possible that this never made it too far beyond the design document.

Garfield itself, in case you are reading this in a place where the comic isn’t published, is a cartoon cat, with the comics having started back in 1978 by artist Jim Davis. Garfield loves lasanga, and making life…interesting for his owner Jon. A number of games and cartoons have been made about him over the years, but when Garfield’s Quest was being proposed at Atari, he was merely a comic strip character who was just getting started in showing up in things like TV commercials.

Who was working on this game is unknown, as the authorship of the document goes by the “Garfield Development Team.” If I were to guess, maybe this was being done by the team who had been working on Crystal Castles, which had launched in July 1983, but that’s just my conjecture. Reading the basics on gameplay though, it does mention that the playfield is made up of “a variety of fixed point perspective views,” so perhaps it was being made like Crystal Castles or Congo Bongo. For the basics of gameplay:


Garfield is a medium resolution (512x384 pixels) color raster game with
the horizontally-mounted monitor showing a variety of fixed point
perspective views of the playfield.


The ultimate objective of the game is to help Garfield locate and rescue
Pooky, Garfield's intimate companion and teddy bear.  During the course of
the game, other objectives will be presented to the player, with
appropriate risk/reward tradeoffs. 

The game would’ve had three chapters, with four “scenarios” in each (the document writer misspelled the word). Those scenarios that make up the gameplay: A cut-scene to setup the situation of the chapter(it doesn’t call them cut-scenes, but that’s what it describes); A Cue-Response scenario (sounds like a Quick Time Event to me; Dragon’s Lair had been released in June 1983, but the term QTE hadn’t been coined yet); A Continuous-Action scenario (playfield scrolling & Evil Otto mentioned); and an Exploration scenario, where you’d look for clues to find the burglar. The document then lays out the chapters+scenarios as to their themes:

I. Urban (Neighborhood, City)
		0. The Abduction		(animation)
		1. Ride'em Cowcat		(cue-response)
		2. Encounter in the Park	(exploration)
		3. Fantasy in the Supermarket	(continuous-action)
	II. Rural (Countryside, Farm)
		0. Hitchhiking to the Country	(animation)
		1. Down on the Farm		(continuous-action)
		2. Bathtime			(cue-response)
		3. Escape			(exploration)
	III. Wilderness (Woods, Mountain)
		0. Into the Woods		(animation)
		1. Lost in the Woods		(continuous-action)
		2. Bobcat Territory		(cue-response)
		3. The Hideout			(exploration)
	IV. Twilite Zone

The rest of the document, that you can read at the AtariAge link, then details ideas behind these scenes and how they might have worked…the cue-response portions do sound like they would have required some guess work on the player’s part, putting it into the realm of trial-and-error that you often had with early QTEs. But, it wouldn’t have been all QTE like many laserdisc games, so the other scenarios probably would have been welcome in that regard. Overall they had some variety in the gameplay, with it starting over at the first scenario if you managed to make it to the end and rescue Pooky.

One interesting note near the end of the document delves into the hardware that they wanted to use. Had this game been produced, it would have been the first arcade game to have employed a hard drive for data storage. The paragraph quoted below touts all of the benefits that using a hard drive would have given the game, including the ability to retain data without the use of also-expensive NVROM chips. Of course, this is something that could have sunk the project, as hard drive prices in 1983 were not at all what they were today, and this being the Crash Era, Atari would end up cancelling a whole slew of projects that would end up costing too much (The Last Starfighter Arcade being one of those):

This game is intended to be a showcase of animation.  To achieve this, the
game will utilize a 10Mbyte Winchester disk drive to store graphics
material for the display.  It is anticipated that the drive can
store up to 200 backgrounds for playfields, or up to 1000 pictures for
motion objects (or some combination of the two).  In addition, the drive
could be used to store program code and data (wave tables, characters and
backgrounds data base) to reduce the memory requirements for the game,
replace non-volatile ROM for high-scores, options, and histograms, and
collect statistics from play of the game (for field test data).

The game hardware probably would have been classified as Atari’s 6502 Color Raster platform, although the precise configuration would change from game to game. Also from Atari Age, here are the hardware specs:

- T-11 Microprocessor with 32K bytes of program ROM/RAM, 4K bytes RAM  (main)
- 6502 Microprocessor with 48K bytes of program ROM    , 2K bytes RAM  (aux)

- 128K bytes of DRAM
- Shared by Motion Objects and Playfield
- Holds 1024  16 x 16 stamp pictures (any MO/PF mix)
- Loaded directly from Winchester Disk

- Scrolling playfield (medium res)
- Displayed playfield  384 x 512  - 24 x 32 stamps (16 x 16)
- Complete playfield  512 x 1024  - 32 x 64 stamps
- 4 bits deep (plus 2 palet, 2 priority)
- H & V reflects on a stamp basis
- Up to 1024 stamp pictures
- Independent H & V scrolling

- 41 motion objects (from 256 specifiers)
- 16 x N (programmable height of 16, 32, 48, 64)
- 4 bits deep (plus 2 palet, 2 priority, 8 bit link pointer)
- H & V reflects
- Up to 1024  16 x 16 pictures

- 48 x 64 stamps (8 x 8)
- 2 bits deep (plus 2 pallet)
- 256 stamp pictures stored in 4K bytes of ROM

- 16 bits of color (4 red, 4 green, 4 blue, 4 intensity)
- Playfield     	  4 pallets x 16 colors
- Motion Objects	  4 pallets x 16 colors
- Alphanumerics 	  4 pallets x 4 colors
- Prioritization for playfield and motion objects

- Playfield		16K bytes  (64 x 128 x 2)
- Motion Objects	 2K bytes  (256 x 8)
- Alphanumerics		 6K bytes  (48 x 64 x 2)

- 10M bytes formatted capacity
- DMAs directly into program, PF/MO DRAM
- Maximum transfer rate of 246K bytes/sec (32K byte blocks)
- Error checking and recovery
- Controlled by T-11
- Disk replaces EEROM

- Quad Pokey (8 channels)

So that helps give us an idea of what could have been. What do you think, does Garfield’s Quest sound like the kind of game you would have wanted to play, or is it better than it never got out of the paper design stage?

Leave A Response »