Every one of you owes it to yourself to read The Art of Prolog. If you liked On Lisp, SICP, or Paradigms of AI Programming you will love it and it will change how you view programming. And, as Alan Kay said, point of view is worth 80 IQ points.
Absolutely. While some of the other introductions to Prolog cover the language quite well (_Programming in Prolog_ by Clocksin & Mellish is good), _The Art of Prolog_ unapologetically dives headfirst into the deep ideas in logic programming, then comes back to the surface and shows how standard Prolog captures them. Its focus on the big ideas ties it less to Prolog, and it sits very comfortably next to books such as SICP and CTM.
It's out of print and a bit expensive ($80-130ish), but the first edition goes for much cheaper. Prolog standardized between the two editions, and the second edition has 100+ pages of new material (including projects), but the former (closer to $6) will give you a pretty good taste.
Well, new copies are available from Amazon in the US, etc. as well, it's just a bit pricey because it was last printed in 1994. (There's not a newer UK edition.)
As an aside: There is a really good section on Prolog in Chapter 7 (Logic Programming) in CTM. Peter Van Roy is quite familiar with Prolog (including several published articles about efficient implementation), and he includes a good summary of its strengths and weaknesses. The chapter on amb/backtracking in SICP doesn't really do Prolog justice, IMHO.
Wow this brings back memories. I studied this language in europe back when I was a kid (must have been around 20 years ago). Very interesting and it definitely changes the way one thinks about programming. Too bad the purpose it was written for (the so called expert systems) never really became popular.
> Too bad the purpose it was written for (the so called expert systems) never really became popular
My impression is that expert systems became too popular (for a while, they were an overworked subject in CS academia and industrial hype literature alike) and died from an excess of broken promises.
For a current successful commercial expert system see for example http://www.zenprise.com -- Zenprise is a rule-based expert system that monitors Microsoft Exchange and Blackberry server deployments, and Active-Sync (Windows Mobile, iPhone, etc.) and Blackberry phones. It automatically monitors system configuration and real-time telemetry to find where problems are in a corporation's message system and help administrators avoid those 3 a.m. phone calls from irate CEOs who aren't getting their emails while working/traveling in Asia.
There are several thousand rules actively monitored and rule violations trigger alerts with resolution plans ... IT personnel can go through the resolution to further troubleshoot and fix their problems. It also identifies the underlying cause for a particular problem (root-cause analysis) ... when one link goes down there are many symptoms. The system's set up to identify the most likely common cause for all those systems. Pretty nice.
A pretty minimal introduction to the language; it doesn't much discuss why one would prefer programming in Prolog to writing the same things in Erlang or Scheme. But within its scope it's pretty fair.
Well, prolog is especially good at a certain task. And that is storing and dealing with logically organized language like data. Back in the time, people used to refer to this as storing and dealing with knowledge instead of data.
The main idea of prolog is to kind of mimic the way the human mind stores knowledge. So eventually one would be able to store the knowledge obtained by an extpert, such as a botanist, engineer, car mechanic, attorney, etc. and be able to recall it or use it to solve problems in simple language like queries.
That is a pretty ambitious goal and, to the best of my knowledge, I don't think it has ever really worked. Nevertheless, it is a very cool language and its ultimate goal may work now that computing power is so high.
