#016x (Gaiden 01) - Animal: "Machine Learning" Taken Literally

Release Date: April 6, 1973

Platform: Mainframe

Genre: Non-game

Developer(s): Arthur Luehrmann, Nathan Teichholtz, Steve North

Publisher(s): Digital Equipment Corporation

And now for something... really different. I'm not even sure this qualifies as a game, to be completely honest. It's more like... artificial intelligence training. 

I'll explain what I mean. Animal is a guessing game, except the one doing the guessing is not the player, but rather the computer. Your role, as a "game master" of sorts, is to teach the computer new animals. You start by thinking of an animal, and then the computer will asks questions to try and guess the animal. It only knows two to start with - fish and bird - and only three questions, but you can teach it new animals, and to ask questions to distinguish between animals. 

The concept for Animal was developed at the home of BASIC, Dartmouth College, by Arthur Luehrmann. I found a substantial interview he did with Kevin Bunch at Atari Archive back in September, 2022 that gives plenty of background on Luehrmann's life and career. I rather enjoy a lot of Bunch's work, and it's a great interview. Luehrmann was a physics professor at Dartmouth from 1965 - 77, and was an early champion of BASIC, writing a few games other than Animal, including one called Potshot. Animal is briefly mentioned in the interview, but is not the main focus of the interview. Most the discussion they have on that game is on the need for a "filth filter" to parse out all the possible vulgarities college students might invent when given free reign to type whatever they like into a game. Animal just seemed to be another of the many experiments conducted with BASIC during its formative years.

I hope that's not how the computer pictures said animals...

Digital Equipment Corporation later got their hands on Animal, where it was modified by Nathan Teichholtz and included in 101 BASIC Computer Games. Steve North of Creative Computing (David Ahl's post-DEC organisation) would further modify the game when the microcomputer edition of the book released in 1978. One of the distinguishing features of the original game was its being one of - if not the very first game - to have a save feature. All of the information you taught the computer could be saved to be used next time the program was run. Theoretically, you could spend a couple of hours teaching the computer several animals, and then invite a friend to test what the computer had learned.

However, the save feature is not in the 1978 modification of the game, which also happens to be the version I have access to. The write-up claims that you could theoretically modify the game to allow for a form of saving, but that could only be done if your system allowed for it. Admittedly, I think this kind of defeats the purpose of Animal. You need to spend a decent amount of time teaching the computer various animals, and having nowhere to save that data spoils the fun.

The '78 edition page on Animal. Now with less scary monsters.

I know I've referred to Animal as a game throughout the article, but, after researching and playing it, it's not really a game at all, if I'm being blunt. My thoughts circulated into one summary point to describe Animal - and it's a weird one, so bear with me - which is that the program is essentially a flowchart creator. The computer asks you a question, like "does the animal swim?" and, based on your answer, it will ask another question (once you build up its database.) It's literally:

• Ask question
• If yes, then A
• If no, then B
• Repeat until animal is guessed

And you can expand the flowchart as much as you want. It's a novel idea, and, as the '78 edition of BASIC Computer Games suggests, it has potential use in an educational environment.

What? I like otters. They're cute.

All this is to say, that I'm making this my first Gaiden article. To recap, in case you haven't read the "My Process" page linked in the sidebar (which is probably due for a revision,) I have two main categories of articles: regular articles and "gaiden" articles (Japanese for "side story" - I shamelessly ripped this idea from Fire Emblem.) Regular articles are the main article, denoted by a number, and are simply all the games that my rules permit me to apply a score to. Gaiden articles are marked with an "x" after the number of the previous article, and are for games (or software, in this case) that I cannot in good conscience provide a score for. Usually this would be for multiplayer-only games, but also the odd non-game like Animal. They're also not counted in the game statistics in the sidebar.

So, I won't be scoring Animal. It's not a game - at least, it doesn't fit within my definition of a game. It's more of a historical curiosity; a glimpse into the infancy of artificial intelligence, and also a novel flowchart creator.

Don't forget - if you enjoy my blog, be sure to leave a comment and follow so you don't miss any updates!