Saturday 11 March 2023

"Doctor Who Adventure" Jeremy Guggenhiem (1983)

In a 2000 post, Jeremy Raynor claimed that this is the first ever Doctor Who computer game. It took the form of a 3 page of listing in the March 1983 edition of the British magazine Computer and Video Games. Like Jeremy I typed it in.  However, as I did so I converted the code to Micro Color BASIC from the original Atari 8-bit BASIC code. The original version seems to have had a creeping memory fault, which eventually corrupts the redefined character set, used in the game for all the onscreen graphic objects. Jeremey asked if anyone out on the Net, who was good at debugging BASIC, could figure out the problem. I think I have.

There were lots of un-RETURNed GOSUBs. A central room movement handler routine called by the main loop would GOSUB to the various different room description and character movement subroutines. But then sometimes Guggenhiem would use a GOTO to return back to the main loop, other times the room handler would be called *again* by room character movement routines when the player left those rooms. This would leave GOSUBs uncompleted or add further GOSUBs (recursively using the room handler jumping routine again). The stored memory addresses for each RETURN would pile up in stack memory somewhere.  Eventually that stack would accumulate enough unhandled addresses that it would begin to cannibalize the area of memory used for the redefined characters. Guggenhiem himself was aware of the problem and in the original article said it could occur after "15-50 games," but he couldn't figure out what caused it.

It is strange that it would occur across multiple games because Guggenhiem uses RUN to restart the program at the end of each game.  However, he also redefines some variables at the end before issuing the RUN command so I suspect that Atari BASIC doesn't necessarily clear all memory and variables back to null when a RUN is issued. Perhaps some special clearing command needs to be issued to get a true cold start. This is different from Microsoft BASIC where RUN reinitializes pretty much everything (string spaces, numeric variables, FOR loops and their return address, GOSUB calls and their addresses,  etc.)

I basically had to resolve this problem by getting rid of the use of GOTOs to jump back to the main game loop.  Instead, what I did was create a kind of pseudo "recursion" feature in the main program loop which calls the room handler (itself a subroutine).  When the room handler routine GOSUBs to a different room routine it increments a counter.  Then instead of jumping directly back to main loop, if that subroutine calls the room handler routine again to go to another room, the counter gets incremented.  An IF statement checks after each call to the room handler routine.  If the counter is not zero a RETURN is issued.  This will return the program to the room handler routine.  That routine then decrements the counter and issue a RETURN (which takes it back to the room handler again) until the counter is zero.  At that point a final RETURN is issued returning control the main loop, which had made the initial call to the room handler.  At that point all the RETURNs have been cleared, and a fresh call is then made to the room handler routine to put the player into the correct new room.  New calls might be made from that routine (an attempt at recursion) but each time a return is issued a cascade is triggered eventually returning control to the main loop, allowing for a fresh call to the room handler and the next desired room.  All this had to be done to eliminate the use of GOTOs and prevent any buildup of unresolved recursive calls to the room handler subroutine.

I made a few changes to the game to make it more in keeping with my understanding of the character of the Doctor. The original game called for the use of a ray gun to kill the Master. My understanding of the Doctor is that he never uses violence (God love him).  So I changed it so that the ray gun is "broken."  The player must figure out what to do with it so that the Master will end up trying to use it against the Doctor, resulting in its malfunction, which will kill the Master by his own hand.  Similarly, the player must avoid the deadly snake rather than kill it and find a way to ward off the deadly Sphinx rather than kill it.  I've added one red herring item to slightly increase the complexity of the puzzle.  Mishandling that object adds to the dangers that the Doctor' must face in the already dangerous maze/pyramid he is exploring.  I also added a rogue/lost Dalek to the monsters that must be faced.

Of course, instead of character graphics as in the original, I had to create new semigraphic images using semigraphic 4 characters. These take the form of 2X2 character blocks (4X4 semi-graphic pixels), which makes for a 16X8 map screen.  I think the Atari game used some kind of double-width character mode, because it seems to run on a 19X21 grid, so the difference between the two screens is not too great. The countdown clock is just a decimal variable decrementing counter used by all routines.  I had to shorten some of the messages because of the smaller screen of the MC-10.  Messages are printed in the "wall" spaces, since there are no spare lines at the bottom of the screen as in the Atari version. I used Greg Dionne's key PEEKs to get 8-way movement from the AWSZ keys used singly or in various combinations with each other.

Contrary to what Rayner suggests, this is not the first Doctor Who game. I have converted a text adventure by James Smith from 1982:

But it is perhaps among the earliest. It's a fun little graphic puzzle game with some minor arcade elements.  Actually, I find it hard to characterize its genre exactly. I wonder whether it is an early example of a hybrid genre of graphic-puzzle-adventure games.  There is definitely a story, including a detailed backstory from the original article, but also the story unfolds in the game itself as the puzzles are solved and the graphical "dungeon" (pyramid) is explored. So I didn't know whether to put it in my "Arcade", "Text Adventures" or "Puzzles" subdirectory.  But I eventually decided all the movement and visuals, and need for some quick reactions make it more of an arcade-style game. Certainly the use of the joystick in the original speaks in favour of this conclusion. In any case, here is the directory where the source can be found:

https://github.com/jggames/trs80mc10/tree/master/quicktype/Arcade/DocWho

The game be can played online using a TRS-80 MC-10 emulator here: https://archive.org/details/@james_gerrie

No comments:

Post a Comment