Spelunking in Colossal Cave Adventure’s Source

In the Digital Humanities Quarterly Dennis Jerz of Seton Hill University wrote a phenomenal essay, in which he explored both the source code and source cave of “the classic text computer game Colossal Cave Adventure”. ~

Adventure was an early video game that circulated on the pre-web internet that was hugely influential for later interactive fiction. Interestingly, the game was originally made by Will Crowther and later modified by Don Woods. One of Jerz’ goals in the piece is to settle cases of authorship, which proved difficult when dealing with fourty year old memories, but was clarified by examining the source code. Jerz also explores the actual cave that Crowther based his original game map on, finding that many in game locations exist exactly as described.

In his study, Jerz explores the source code of Adventure before and after it was modified by Woods. These files were found on a backup of Woods’ old Stanford student computer account, an interesting find for digital archeology. Jerz determines that, despite common claims to the contrary, Crowther’s original game was not solely a cave simulator. It contained fantasy elements, puzzles, and some of the magic words that entered broader computer culture, such as “XYZZY”.

Jerz spends much less time on the source code than he does on the source cave. Crowther was an avid caver, basing the geography and some in game items on the Bedquilt entrance of Colossal Cave in Kentucky. For the essay, Jerz was led through the cave by several old caving companions of Crowther, and he includes numerous photographs of locations in the cave that correspond to the game. By combining the source code, cave, and interviews with Crowther, Woods and their family and friends, Jerz is able to construct a fascinating study of the authorial process of a historic early video game.

While very fruitful for the study of Adventure, Jerz approach is not the way forward for reading code. I would hazard a guess that most games owe more to untraceable imagination, or inspiration from other media, rather than a specific physical location. Indeed, Woods had never been caving, and he introduced puzzles that broke from the experience of an actual Kentucky Cave. Jerz did not intend his study to be “Critical Code Studies”; indeed no such label existed when he began the article. Still, his technique of explaining the workings of the code with relation to the surface level of the game is one I hope to emulate.

Adventure is actually a game whose source code has experienced a relatively large amount of critical study, when compared with most games. It was the “sole example” provided by Donald Knuth of his CWEB system for literate coding. After Jerz article in DHQ, the Critical Code Studies working group explored the code further.

The discussion ranges across the board, from cultural studies of the 1970’s programming subculture to personal recollections of scholars’ childhood experiences with the game. Many people commented on the sheer difficulty of reading this code, which is written in a computer language and paradigm that is incomprehensible to those of us who have learned to code post-“goto considered harmful”. As one clever slashdot commentor noted “Maybe the inspiration for the ‘twisty little passages, all alike’ wasn’t Mammoth Cave, it was the code itself”.