My Life With Zork

In addition to getting some background on HCI in yesterday's class, we also discussed early computer gaming. Our assignment was to find an online version of Zork and play/map it for an hour or two. I made an off-hand remark something like "I've put more hours into Zork than some of you have been alive!", which got a few chuckles and some patronizing glances.

But I was just being truthful!

When I first started computing (with an Apple //e), Zork was among my first purchases. I had seen Adventure/Dungeon on minicomputers (DEC systems) in the past, but had never played one personally. I was hooked, spending hours upon hours mapping my progress, trying different flavors of verbiage and trying to get through those "twisty little passages" to finish up the game. I got a hint from a local Apple bulletin board (does anyone remember those anymore?), and finally put it to rest. Of course, that led to the purchase of Zork 2, Zork 3 and a ton of other Infocom games - all text based, and all very, very difficult.

Often, when people talk about the Zork series, they talk about the primitive user interface and the fact that it seemed to be "stolen" from the public game "Adventure". What they don't understand is that the innovation in these games wasn't about the story (although Zork 2 and 3 extended the story extensively); rather, it was about improvements in the text parser, and trying to get to the point of natural language commands.

At the time, everyone thought that natural language would be the be-all and end-all of computing. If we could only talk to the computer like they did in Star Trek NG, we would be saved! Each version of Zork had a better parsing engine, and the computer mags would breathlessly proclaim the vast language and syntax that each new version contained. Alas, the result of this effort was Infocom's Cornerstone database, which tried to put a natural (Zork-influenced) language in front of a database system. I was one of the four people that bought that thing (which led to Infocom's demise, really), and I never was able to make sense of it. Perhaps the most useful thing it its (very attractive) box was a "Don't Panic!" button that told of the software's heritage. It was good advice.

Alas, at some point the natural language front-end was proven to be pointless, leading instead to visually-accessible databases (led by Reflex, one of my favorite software packages ever) and proto-programming interfaces for databases (SQL, primarily). But an interesting stage of computer history, that's for sure!

[ddg]