FYP

[Anonymous]

Hey everybody! :D

I’m entering 4th year this year and still undecided about my FYP. My major interest is in programming and software development and though I want to get into gaming, I do not want to limit myself to this. So, does anybody have any ideas? Anything appreciated :wink:

[Anonymous]

Colkorsu,

I guess much will depend on where (and what) you are studying, resources available, etc.

If your doing Comp Sci (or equivalent), and a games job is your end goal, then try to think of the specialised type of programming job you would like.

AI is usually a good one to start with, DFA and nDFA are always good places to start, maybe steering behaviours and/or locomotion models. Squad-based AI is also a pretty interesting topic.

Physics is a bit harder, unless your a physicist, but with Havok and PhysX both available now, with docs and forums, its not impossible.

Mathematics-based projects, such as partitioning algorithms, hybrid scene managers, are always welcomed by academic panels.

Graphics. Well, DX10 is out now, with a fully open pipeline, so getting stuck into Cg or HLSL for a year might be pretty fun. It will be pretty intense, and of all the above, I think effects programming is probably the hardest. (in my opinion).

Networking. Sure. Api’s like RakNet, and the success of Demonware, and emerging topologies like Microsoft’s LIVE, would give you a wide scope of topics. And if networking is something you have done in years 1,2, or 3 of your degree, then this may be pretty straightforward.

I guess at the core of any of the above (or a mix), you should have a clear point of difference, or innovation. You will need to find a current constraint or challenge, that exists in the field, and then research and develop methods to solve that problem. Depending on where you are studying the focus may be on the actual prototype(s), or possibly on the research document.

You should have a look at your lecturers research background also, as if you can find one who shares a common interest, it will definitely speed up the process involved in identifying the "state of the art."

Also, think about who you want to work for. Specific companies, their titles, genres, etc. Read their "post-mortems" if they are up on GamaSutra, and see if there is any ideas there.

Just my 2 cents.

Brendan.

[Anonymous]

Thanks for the detailed reply Brendan,

Well i’m on the Comp Games course in UL, which is pretty much the Comp Sys course with few extras, but we are not restricted in our FYP.

I want to do a mathematical based project. My projs thus far have been mainly graphics based so maybe AI or collision detection algs would be good..

Now I just need to choose and get on with it eh! :?

[Anonymous]

Hi Colkorosu,

check out the IGDA’s special interest groups here:
http://www.igda.org/SIGs/

plenty of papers and presentations relating to the state of the art.

B.

[Anonymous]

Hi guys,

Personally, I think there are already a lot of FYP’s that cover AI, pathfinding and collision detection.

If you’re good with graphics, and want to do something a little more mathematical, why not try something SpeedTree like?

F

[Anonymous]

Hi KookiMonster,

Actually, strangely enough, I was planning on doing a fractal system for waves last semester as a side project but it never materialized. I did a bit of research on fluid dynamics at the time and found some really impressive stuff, like http://www.frankpetterson.com/index.cgi

I got a little intimidated so…
Would this be too much work to write everything from scratch? Or should it employ a rendering engine?
I would prefer to use an engine of some sort so I wouldn’t have to write scene manager etc.

We have one lecturer who’s really into Graphics so i’ll go have a chat with him and see what he thinks.

Cheers man! Any more info on that subject appreciated!
Colin

[Anonymous]

Would this be too much work to write everything from scratch? Or should it employ a rendering engine?[/quote:22d8c149fb]

I’d say the rendering engine would be the least of your worries for a project like that. If you use something like SDL or the likes then you could get a basic renderer up and running pretty quickly.

The really hard part would be getting the model for the water behaviour right, and doing it all at acceptable speeds. I don’t know if you want this stuff to run in realtime or as an offline process, but If you’re aiming for realtime you might want to look into NVIDIA’s CUDA api or ATI’s Stream to help leverage the grunt of the GPU for the task.

[Anonymous]

Colin,

You should look at your FYP as your all in one portfolio for getting a job – so the more you do yourself, and the more impressive it is, the better. At the same time, you want to focus the most effort on the core of your project; if you’re doing water, spend most of your time on the water. Same for trees, or grass, or clouds, or whatever.

Check with UL and see if they have licenses for any engines – if they do, use one of those. Anyone looking to employ you will want to know that you can work within their existing system (engine, toolchain, etc) with a minimum of fuss, rather than insisting on wasting time doing everything from scratch (unlike a student I once had who refused to use STL, because he could write better versions himself).

F

[Anonymous]

Colin,

You should look at your FYP as your all in one portfolio for getting a job – so the more you do yourself, and the more impressive it is, the better. At the same time, you want to focus the most effort on the core of your project; if you’re doing water, spend most of your time on the water. Same for trees, or grass, or clouds, or whatever.

Check with UL and see if they have licenses for any engines – if they do, use one of those. Anyone looking to employ you will want to know that you can work within their existing system (engine, toolchain, etc) with a minimum of fuss, rather than insisting on wasting time doing everything from scratch (unlike a student I once had who refused to use STL, because he could write better versions himself).

F[/quote:956cd9fc92]

Solid advice, though I’ve give some very careful thought on whether you want to use an engine or not- depending on what your aims are.

If you’re only going to be drawing some particles or a water surface contained in something like a jug or a bathtub then employing an engine might actually be overkill- it’d be like taking a sledgehammer to a nail.

On the other hand if you want a more complex scene than that then it might be more beneficial to use an engine.

Engines are not a free meal and it does take time to become accustomed to them (and their limitations) so you need to weigh up the time it would take to do what you want to manually versus the time it would take to learn the engine and get your stuff up and running in it.

Just my own two cents.

[Anonymous]

Thanks for the input!

I’ve decided to continue on with this project. I wont use an engine but I want to focus mainly on the math. In the end of this I will make a relatively small demo of the underlying work.

Looking forward to learning more about fluid dynamics! An interesting topic.

[Anonymous]

Thanks for the input!

I’ve decided to continue on with this project. I wont use an engine but I want to focus mainly on the math. In the end of this I will make a relatively small demo of the underlying work.

Looking forward to learning more about fluid dynamics! An interesting topic.[/quote:118a7d8ead]

Best of luck with the project! :) Doing a FYP myself this year myself so I know how difficult it can be to decide!

If you’re interested also I’ve got some OpenGL framework code that I reuse a lot between assignments. It handles all the mundane stuff like setting up windows & displays, loading textures, doing basic text rendering, and so forth. It’s pretty straightforward to use and might save you more time to concentrate on the physics.

Let me know anyhow if that would be any good to you and I can send on a simple example where it was used.

[Anonymous]

Darragh,
Thanks man, you too! That OpenGL code sounds good if you can get it to me! Before I used Obj-C wrappers for the GL code but want to steer away from Mac dependencies to move more cross plat. :P

Thanks a lot

Colin

[Anonymous]

No problem your welcome!

Here’s the little project where it was used anyhow:

http://www.darraghcoy.com/storage/PhysicsPractical1.rar

It’s a simple physics practical that was done in 3rd year. Just a rotating turret/gun that fires a cannonball at a crate. The main files that would be of interest are PhysicsPractical1.cpp and Main.cpp, the rest is all non-application specific stuff.

The code includes a project for and compiles with Visual C++ Express 2005, but should work in 2008 as well without too much trouble. GCC shouldn’t be hard to target either since I used to use the GCC compiler up until not too long ago.

If you’re using VC2005 you’ll need to download the Windows Platform SDK in order to be able to compile the code, since the native (Win32) Windows libraries are not included with VC2005 as standard. I don’t think that is the case for VC2008 however (anyone know for certain?)

The Windows sdk (for VS2005) is available from here:

http://www.microsoft.com/downloads/details.aspx?familyid=D8EECD75-1FC4-49E5-BC66-9DA2B03D9B92&displaylang=en

Once you’ve downloaded and installed the SDK, setup visual studio to use the include and library directories in the SDK install directory by going to:

tools->options->Projects and solutions->VC++ Directories

After that’s done Visual Studio should recognise the Windows libraries and the project should compile. If you’ve any more questions then send me a PM or an e-mail and I’ll be happy to answer them.

[Anonymous]

Hi Colin,

Fluid Dynamics is a very interesting (and large) field. I have a paper and some example FD code that we used in our last game if you want it?

F

[Anonymous]

Thats a nice code-base you have for yourself Darragh! Good example of software reuse eh!
Will come in handy thx! :)

K,

Ya that would be excellent if you can! :D Gathering some research on it at the moment so the paper and/of code would be useful. Whatever you can would be great!

[Author]

Posts