Ok peeps.

I am in a group of 6 people for my final year university project. We have to develop something, what to develop is up to us but it has to have a suitable workload and "academic" value (the program itself doesn't have to be academic, but the more academic hoops we jump through when developing it the better).

My project last year was an environment to configure and run cellular automata and it got a good mark, but as a software project it was quite basic. So if you multiply that workload by about 7 you'll hit what we need to do.

Problem is that we don't actually have any ideas. The rest of my group is under the delusion that despite having minimal (zero) c++ and opengl/sdl experience we can write a multiplatform game engine (or development library) in the timeframe (september to april). We could probably make a demo of a game (as long as we could find suitable academic merit), but not a generic engine.

People generally have interest in 3D graphics and AI.

Any ideas? (you don't have to keep it to graphics and AI if you don't want to, but such ideas have most chance of being accepted by the group). This team is really (rubbish) in trouble and my ability to make them use some common sense is quickly fading.

Please help we really need practical ideas...

A program that makes you less British. Somehow.

I could see a small, easy game being made. Maybe a roguelike. Really, making a game isn't THAT hard, especially given 7 months' time. My friend made a decent one in about 4-5 with SDL (C-based GUI programming), though he's still adding stuff, it's functional and kinda neat.

Making a game isn't that hard (depending on type of game), making an engine IS. Especially if said engine is to have academic merit rather than just being another one of the hundreds of freeware engines out there.

Coming up with a game idea that university staff accept as a good project is quite hard though.

Oh yeah, we actually have to COMPLETE the project.

I was thinking AI before I was halfway through your first post, but then you later mentioned that you need to complete it :p

Black hole calculator.

XD

You guys could write your own interpreted programming language with super psychadelic features

idea... create a new audio compression format...

that kind of stuff was 3rd year (solo) project material.

rofl Det. Okay, sure...

Then write an operating system all in assembly that outsells Windows by April. Challenging enough?

Compression formats were 3rd year projects because 3rd year projects didn't have to be finished, and it was the kind of thing that could be done by one person if the idea was good enough.

I'm glad you made note of the final line of my first post though, about looking for feasible ideas.

I'd rather you didn't contribute at all than bring stupidity into my thread.

<3 Det.

What about a Graphical graphing calculator?

Actually, an operating system sounds pretty cool, though I'd write most of it in C, which doesn't seem like something you guys are familiar with.

Do you guys have any other interests? Does it have to be a lot of code or just a lot of work? Doing research for a professor could be a good idea (which is what I did for my senior project). It seems like your best bet is an AI project, though there are way too many to choose from. You could make some kind of expert system or a complicated (but not necessarily graphical) game.

I think an interpreter would be too easy, especially with lex/yacc. I can make a basic calculator in five minutes and a decent interpreter in a day or so. I did have to write an interpreter in C++ before (sans lex/yacc) and it wasn't too bad. It would have to be pretty feature filled or unique to be worth it.

I'd do something creative with neural nets. I wish I could give you a clear idea on what to do with them, but unfortunately, I can't. As you may have heard, things like image recognition, adding and subtracting, can be accomplished via a neural net by teaching it "facts." If you can come up with some sort of task for it to do that would be hard to accomplish via regular code, it could be impressive.

And, on the off chance that you haven't already seen the following articles, I figured I may as well include them:

Generation5 Neural Net Articles

Create a fully integrated media frontend for any given OS (like Media Center for or Meedio for WinXP, only better) with fully integrated media player (supporting most known formats including DIVX6). I'd do it myself, but my complete and utter lack of knowlege on the subject of coding makes it nigh on impossible.

Basically we have to make something that can be viewed as a product. It's for the engineering accreditation in my MEng Computer Science degree, so it's about project management and making a finished product. Which means:
- No research projects
- No project for which the success relies on something we may not even be able to achieve (like an audio compression algorithm better than anything that already exists)

Mmmmm...

maybe some kind a 'learning' macro type program,

like (while its running) it notices when I use combos of programs at a time it would open all of then when I start one, like when playing JA, when I open JA it would know to open TS xfire,minzser,/ect

and when I start my computer and open firefox it would learn to auto open the sites I cheack very time I start the program

ok I have no idea if thats possable or if its been done allready, but just an Idea..

What about dealing with something on the more hardware side? You could maybe use some recent technologies such as bluetooth, RFID, wifi, etc. Granted, I don't know what you could do with them.


Another thing you could do is to make some sort of real time, multiuser 3d world. Yeah, its been done, but add an element to it that makes it unique. Something like adding the ability to add objects to the world in realtime via a 3d scanner. I know you said practical, and a 3d scanner isn't that hard to make (http://splinescan.co.uk/index.html). Anything that connects a 3d world to the real world would have to be of some academic interest.

Anyway, just some suggestions.

On topic of games, why not make one of those 64KB graphical demos?

[QUOTE=Grim Zombie]Mmmmm...

maybe some kind a 'learning' macro type program,

like (while its running) it notices when I use combos of programs at a time it would open all of then when I start one, like when playing JA, when I open JA it would know to open TS xfire,minzser,/ect

and when I start my computer and open firefox it would learn to auto open the sites I cheack very time I start the program

ok I have no idea if thats possable or if its been done allready, but just an Idea..[/QUOTE]

http://expect.nist.gov/

Autoexpect actually does something similar, well, in the command line anyway.


Why not just write some practical application, like a paint program or a word processor? Or does it have to be really unique?

I was actually just today looking for some projects for my Software Engineering class. Our group decided to do an inventory control system for one of the audio/video labs on campus that is still keeping track of their schedules and equipment with paper, but while I was looking around, I stumbled upon a couple of really neat ideas.

Most of them have to do with fulfilling some of the bounties that open source projects such as Gnome and Ubuntu are offering. I'll repost a couple of the more interesting ones:



Creating an equivalent to Apple's iCal in Linux

A distributed grid plug-in for Firefox


Our group considered that one for a bit just because it wouldn't be too difficult to do if we utilized the mediawiki engine, but we decided the scope was a bit much for just one semester's worth of work.

I'm seriously considering this for my capstone project next year, but I'll share it anyway. An Arimaa bot

Solve P=NP. Develop an algorithm that computes Riemann's Hypothesis.

Oh wait it has to be solvable. Well I'm confident that it is solvable! And just think, you'll be $2 million richer!

Crack RSA encryption => prize money => blackmail => profit => retirement.

Alright, how about this.

What about a program that automatically searches out free porn, downloads it, then catagorizes it.

I smell a nobel peace prize on that one.

rofl... but seriously.. that's a good idea.. damn..

There was a CS test question floating around last year that said:

Write a new 3rd(maybe it was 4th) level computer language and use it to write a program that finishes the rest of the test for you.

A COFF execution support patch for the Linux kernel.

how about extending the automata idea - create an environment with rules (i.e. a physics engine, although it needn't necessarily model "real" physics. it can then be scaled in complexity to suit the timescale of the project)

Then create agents to populate the world. These should be learning/evolving agents, so make it a generational simulation and after each generation, the most successful agents get to populate the next generation.

At a simple level, you can have "food" in the environment, one set of "herbivore" agents that eat that food, and one set of "predator" agents that eat the herbivore agents. When I was doing Uni open days, we were told about a similar simulation, and the predator agents learnt after several generations to hide right behind the food, so the herbivores would go to eat the food and then run pretty much straight into the predators.

You can also get some pretty funny effects. They'd given the agents the ability to pick things up and jump. Through a bug in the physics engine, they'd learnt they were able to jump up, and then pick themselves up. That was kinda weird.

that's actually close to one of my ideas.

*pats Giraffe*

And mine.

<.<
>.>

/hugs his MEng