The Prolog Story

http://kylecordes.com/2010/the-prolog-story

by zephyrfalcon • 16 years ago

96 60 16 years ago

This story sort of reminds me of another that was reported here. See the link referred to in http://news.ycombinator.com/item?id=756873

Maro2• 16 years ago

Learning Prolog was one of the most exciting things about getting a C.S. degree. It's vastly underrated, though admittedly not very useful in real-life.

fogus2• 16 years ago

I once worked with a guy who was using Prolog to develop a system that devised optimal cargo hold layouts for shipping barges. Seems fairly "real-life" to me.

Maro2• 16 years ago

I meant real-life applications are relatively few.

barrkel8• 16 years ago

Optimal packing problems are fun 1-hour programming contest fodder - so many problems that you might solve in Prolog are trivially written imperatively or functionally as combinatorial or permutational recursive searches with pruning - but they are such a small kernel of a real-world application that I don't think they justify the complexity of linking in another language, runtime, etc.

Where you're dynamically building up facts and clauses, and you need more general solving logic at run time, then I think Prolog is more warranted. But that's a smaller niche, IMO.

10ren• 16 years ago

For every thousand hacking at the leaves of evil, there is one striking at the root.

jm42• 16 years ago

The book Seven Languages in Seven Weeks has a chapter on Prolog. There is an interview in there with a guy who was doing research on dolphins and used Prolog to build a dolphin simulator. As they learned new things about the dolphins they would add "facts" to the simulator and used it to try to discover new things.

They could also use it for "debugging" training problems. He had a story about a dolphin who would apparently ignore a specific instruction and just sink down into the tank. They were able to use the simulator to figure out that the problem was not with the dolphin, but with both the instruction given and the surrounding environment at the time they would give the instruction. It turned out the dolphin was actually "following the rules", but the trainers failed to see it.

He also talks about using Prolog to build a scheduling system. Access to trained dolphins is a precious resource and it was important to find the optimal schedule for research. With a handful of facts and rules, Prolog was able to discover it for them. Pretty neat stuff.

rick_2047• 16 years ago

do you have any links or resources for this? I would really like to read about this more.

dejv• 16 years ago

I like this sentence: We cut the cost/time of this implementation by 90% or more, not by coding more quickly, but by thinking more clearly.

I think that this is valid for most projects in any language.

sophacles2• 16 years ago

I've always been a fan of the "Don't just do something! Sit there." approach to work. Not only do I sometimes find really clever ways around a problem, or weird issues that wouldn't immediately be apparent, I also save a whole bunch of potential RSI :)

Added benefit: When it is time to start coding, you can amaze your coworkers by brain dumping a decently coded solution at typing speed.

rdtsc2• 16 years ago

That's one of the most understated advice. Quite often managers like to see lots of code, or developers have screwed up incentives and get payed by the lines of code. If a manager walks by, their heart gladdens when they see you banging on the keboard like mad -- to them that equals to "work being done". They fail to realize that sometimes that equals to "more un-necessary garbage added to the existing pile of code" that will eventually make the product unstable, brittle and hard to maintain.

Sitting there with a notepad and a pencil thinking about the problem is when I find I actually solve the difficult problems not when I type furiously at a keyboard.

cadr• 16 years ago

At a talk, someone (might have been Eric Evans) said something along the lines of when he was modeling a domain with a group, he wouldn't let them leave until they had three different ways of doing it. Even if the first was perfect and the last two ended up being stupid, forcing them to brainstorm other ways of doing it often came up with hidden gems.

silentbicycle7• 16 years ago

Yes, but the raison d'être for Prolog is to do declarative/logic programming, which is fundamentally about programming by writing a clear, logical model of the problem, that incidentally happens to be executable / queryable.

"Declarative programming clears the mind." - Sterling & Shapiro, _The Art of Prolog_

Provided you don't abuse the cut operator, a Prolog program will consist of a series of rules which each have a logical meaning in isolation. Reasoning about Prolog code is so different from trying to unravel what's actually happening with several layers of OO polymorphic dispatch, it's not even funny.

"Elegance is not optional." - Richard O'Keefe, _The Craft of Prolog_

tlack2• 16 years ago

Does anyone else have any war stories about using Prolog in their day jobs? It's a tool I'm always on the look out to use to simplify some of my more complex tasks, but so far I haven't had any luck in that regard.

ciex• 16 years ago

I am currently learning Prolog as part of my artificial intelligence class and as far as I know it's used in many contexts in ai research in my uni.

porlw2• 16 years ago

I once had a huge pile of Oracle PL/SQL to take care of, which was written for performance rather than ease of maintenance and used a lot of session level global data.

PL/SQL doesn't give you nice stack traces when you have an error, I just knew the entry point and the place where the error occurred.

Even knowing the data coming in to the entry point, use of the globals meant that you often couldn't know for sure what route the code would take.

I wrote a primitive PL/SQL parser (in Perl) to turn the call tree into Prolog rules, and then used Prolog to give me the possible routes between the entry point and the error location, making debugging much easier.

tlack2• 16 years ago

Wow, that sounds like a serious undertaking. How much PL/SQL are we talking about here? How long did writing the Perl -> Prolog solution take?

MichaelSalib• 16 years ago

At a previous job, one of my coworkers implemented a fairly complex role-based access control system by linking a prolog interpreter to our application. Our problem domain was...very unusual, so this was probably the best solution.

ollysb• 16 years ago

It was the similarity to prolog that first got me excited about erlang, I thought I'd found a language that used the same principles that could be used in the real world. Then I discovered there was no backtracking...

stralep• 16 years ago

Is there some up-to-date implementation of micro-Prolog?

I find its syntax nice and lispish :)

[edit]s/it's/its/

scscsc• 16 years ago

Ok, so Prolog is nice. We know that already. But the story lacks any real content... It's really just: we used Prolog, it seemed to be a lot faster to code and the customer was happy.

mark_l_watson2• 16 years ago

I used to use Prolog a lot on R&D projects at SAIC but since becoming a consultant, few customers have wanted work done in Prolog (it has been about 5 years since anyone has paid me to do Prolog).

Prolog is awesome for some types of problems like graph operations, ad-hoc rule systems, planning, etc.

If you want to play/learn, I would recommend Swi-Prolog. Jan Wielemaker does a great job writing and maintaining Swi-Prolog. The semantic web libraries and other Swi-Prolog libraries are very good stuff!

tom_b• 16 years ago

Thanks for weighing in - I am a somewhat frequent visitor to your blog and websites. This type of discussion (eg, using uncommon or less well-known languages to be more productive or reason about an approach to a problem domain) comes up here on Hacker News frequently.

I have been searching for new work opportunities that would involve more freedom in programming language choice (specifically allowing me, as the developer, to explore tools like Prolog, Lisp, and Ruby as appropriate). Sadly, this has been somewhat fruitless in my geographic region. There is a (probably well-founded) desire on the part of companies to use what we can refer to as "lowest common denominator" tools (Java, C#, Oracle, MS SQL Server) with the occasional upstart using Ruby/Rails. When I say LCD tools, I don't mean to denigrate those languages/platforms, merely point out that companies seem more comfortable knowing that there is a large pool of potential programmers available with those skills.

PG has weighed in on the side of starting your own company and using more powerful languages as a competitive advantage (the whole blub/magic of Lisp set of essays). It seems that few companies see programming language as competitive advantage or (based on your comment) be interested in hiring consultants such as yourself to implement higher level software that leverages these uncommon languages.

I'm curious - have you had projects where you felt that Prolog or Lisp represented a "better" choice than Java or Ruby but had customer requirements that specified language choice upfront? Have you "pitched" Prolog or Lisp to customers, maybe saying something like "ah, using Prolog here would require only 20 hours of my dev/consultant time, but using Java will require an 80 hour solution"?

I'm quite interested in using Prolog, Scheme, or Lisp at the "day job" so hearing any of your thoughts might help better prepare my own internal pitches . . .

mark_l_watson2• 16 years ago

Hello Tom, first I should say that Ruby is my favorite language and fortunately there is a lot of demand for Ruby development. I don't really pitch customers on language since almost everyone I work for is very tech saavy and know what they want they want to use.

If you want to work in a particular language such as Common Lisp I suggest writing a useful open source project and that could lead to work in Lisp, and at the least you will learn something and have fun.

travem• 16 years ago

I've used Drools a fair bit in my current job but I've never learnt much about prolog. Can any one summarise where prolog would work better or provide additional features?

agbell• 16 years ago

Are there things prolog can do that embedded solvers cannot? In .net, solver foundation, http://code.msdn.microsoft.com/solverfoundation , is a great fit for solving from a set of rules. And I believe most languages can embed some sort of rules optimizing engine.

Mark_B• 16 years ago

So, in other words, the customer's system can now be described as being "a client/server application written completely in [not Prolog] with SQL Server behind it except for this ONE piece - THAT one is in Prolog!"

Next question - Who's going to maintain it?

I mean, props for using Prolog to tackle the complex problem and save gobs of money/time, but doesn't an approach like this seem _really_ risky?

dws• 16 years ago

Great takeaway quote: "Don't be the person who only has a hammer."

The Prolog Story

Related Stories