If it works…

…it’s not AI.

Funny. 🙂

Just ran across a great paper written by Eve Phillips documenting a history of artificial intelligence. Well worth a read. The title made me smile because my first job in IT, all the way back to 1989, was writing an expert system and it most certainly did not work.

The above paper reference came from Dan Weinreb’s post Why Did Symbolics Fail?

If you ever want to start a company, you can learn a lot from reading “war stories” like the ones herein.

Alleluia to that, brother…or even if you just want to be a better engineer.

Update August 2013: unfortunately Dan Weinreb’s blog has gone and with it the link to the article. The closest I’ve managed to find is this link on John Wiseman’s blog.

Update December 2013: now John Wiseman’s blog has gone. I think the internet is conspiring against me.

Computer Languages 22 years on

Many moons ago in the late 1980s, right at the start of my interest in computers, I bought a book about computer languages. The book titled Computer Languages: A Guide for the Perplexed by Naomi S Baron inspired me to get into programming and eventually led me to program professionally (in the sense that somebody paid me 🙂 .)

What is interesting, 22 years on, is what has changed and perhaps just as interesting, what hasn’t changed.

The languages highlighted in the book are:

The late 1980s saw one of the periodic booms and busts in artificial intelligence and the languages selected for the book reflects that. Believe it or not PROLOG was considered to be a good candidate to become a mainstream language for a time in the mid-late 1980s. You couldn’t move for articles on artificial intelligence using PROLOG in the hobbyist end of the computer press.

Whilst at college I ran across Intellect, a natural language relational database query tool. It did have a spooky ability to understand your queries so I am a little surprised it never caught on. Sometimes technologies don’t fail because of any technical limitations but more for social reasons. Perhaps database administrators didn’t relish the thought of the great unwashed having access to their precious databases?

The rest of the list isn’t much of a surprise and is a pretty good insight into what the business and scientific community was using at the time the book was written.

The two main language strands not present, because they either hadn’t been invented or were very immature, are new dynamic scripting languages like Perl and Python and object oriented extensions to existing statically typed languages like C++ and Java.

You never know, with highly parallel computing architectures becoming the norm, perhaps the likes of LISP and PROLOG still have a role to play?