What do you think is an interesting programming problem

I am planning to do a contest to promote my website. I want to do a programming related contest.
I'm trying to find an interesting programming problem that can be used as a contest.

Traditionally any typical AI task is interesting enough since there is usually no perfect answer and we can see different approach to different things. But personally I think that all AI problems are pretty much well researched that you can easily find a near optimum solution. I want this to really be a programming contest rather than a who can google better contest.

An optimization problem seems like an good problem. Like design a script to take data from 4 joined tables in the most effective way. But I think that is maybe too hardcore to get a wide audience.

What do you think. You have any cool ideas?

 

There are many challenging theoritical computer science problem as well , else you can ask them to solve some hard to crack puzzle programmatically.

 

Find the last digit of PI.

 

Placing 8 Queen Pieces on a standard chessboard in away that they cannot hit eachother either streight and/or cross. I did this long ago in C++, and found only 3 different placement solutions, becuase of my old dx2 computer at that time.

This would be really interesting.

regards

 

AI is not so well developed as you think, there are many open problems in AI, if you pick some classic problem, it would be heavily studied before.

I think you could pick 2 or 3 problems and give a different set of points to each

 

after some thinking, i'm leaning towards some simple but open problems. Nothing too domain specific. Something that everyone can code. But a good coder can optimize really well.
For example. Given a set of data in multiple tables. Find the best way for user to browse the data.
So some would use one big sql to join the tables to get all the data and query again with every user request, some would use sub selects, some would use multiple sql statments, some would load the whole database in momory and request it from memory. Etc etc..
Something open like that I think will allow more people to join in since you would not need to have any specific knowldge related to a specific AI field.
Would you enter if it was something like that? Whatdya think?

 

Depending on what languages you're looking at, I would possibly try something like a developer framework. Take something like Ruby On Rails (ROR) for example. Numerous frameworks have been written before yet ROR is so popular it's been ported to nearly every web language out there.

The reason I suggest this is because it's good on multiple levels:

1) It requires intensive thought. Implementing different OOP concepts, strategizing the overall handling to support the end-user, and even like you suggested heavy research into performance tweaks.

2) It's very unique. I've seen a ton of programming contests like everyone specified here. Typical AI or increasing performance, pretty bland if you ask me. This requires some personality (you'd be amazed at how much personality is put behind something like a framework).

3) It's creative. I can guarantee no framework would ever truly look the same. The possibilities are endless as to how the programmer can go about making it. I myself am an example when I created PHPCell which is a PHP Framework (yet to be released) that mimics the biological environment of the cell.

4) It helps the community. Who knows what will be produced from the contest. You might get one of the next big frameworks that could move on to be developed by many others into something huge.

I could probably state more, but you get the idea. You'd of course need some basic guidelines like what exactly the framework should help with. For instance if you did PHP you could simply say it should make designing small to medium scale applications easier, and from there the programmer can do what ever he/she wanted to get the job done.

And finally, if I saw a contest like this, I'd be very intrigued to enter

 

Some interesting programming problems that I had solved:
Tree traversal
Create an expression evaluator [like 2+3 +(4/5 -4)=x]
Create some game in some language with max of x Kbytes of size of the source code (or target code).