h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

Hey Henry, I like the look of the computer game you're working on. It looks like you've got about as much done on it as I have on the game I'm writing (can't include a link here, but there's one in my userspace). It's quite difficult to find the time for a maths degree and write a game isn't it?

You've got screenshots of your game - must be further on than galileo
Well, I've been working on the physics of it, other people are doing the graphics. Now that everyone has finished exams and such things, stuff is moving again, hopefully we'll even be able to link the physics to the graphics. We're doing all the physics continuously rather than the usual incremental stuff, so its all new stuff, learning lots etc.

I like this from your website:
"You will be able to enter these with a very powerful but simple to use interface which hasn't yet been designed."

I know that feeling

Ah, the physics engine. I've just started working on that too, bloody difficult (especially for a pure mathematician not used to doing applied stuff). I think I've finally got the collision response system working, but it's hard going.

You have to have grandiose plans (powerful and simple to use ) to start with, otherwise you've got nowhere to slip down to when you start making compromises.

I just read that Black and White has 2 million lines of code in it. Everything I've written for my current game sums to about 11,000 lines! Oh dear, still quite a way to go

Well, good luck with it anyway.

I've been thinking about the physics for (literally) ages, so I've a fair feeling where its going. I'm in the nice position of having other people to write the code for me, I just have to do pseudocode and theorise. This of course leads to very grandiose plans

What sort of physics are you using? Looks isometric viewpoint from the screenshots, even tile based? Can't be if youre writing physics for it. Newtonian motion?

Wow, how did you manage to persuade someone to write the code for you?

The graphics engine is isometric tile based, but there's going to be free and continuous movement on the terrain (like in Tiberian Sun for instance). The physics engine is going to be simple (it's a one man project after all), 2d polygons and circles only (the game is ostensibly 3d, but the spaceships are really hoverships rather than spaceships, so they'll always be a roughly fixed distance from the ground, so I can model it as 2d - sort of), and only the circles moving. Basically, the spaceships are going to have force shields round them, which are circular. The buildings are going to be polygonal, so no movement necessary. It's a bit of a cheat, but I want to actually finish it . As for the mathematical part, just Newtonian dynamics, possibly angular effects as well as linear ones, some friction and restitution effects.

Its quite a large project, been going a long time (but not that much code written...of course...). Theres a number of coders floating around, and they want the stuff to move along, and I don't do much coding though I know the physics theory insideout (worked quite a lot of it out myself).

Is there going to be an effect of the height of terrain in the physics? If youre just modelling it in 2d that might be difficult to incorporate. But things like collision detection will be easy enough.

Yes, the physics is going to be a bit 3d. I'm modelling the way a hovership works by introducing a counter-gravitational force proportional to 1/r^2 where r is the distance from the ground. This, combined with gravity, gives you a natural "hover distance" from the ground, and looks quite cool. If you fly towards a hill fast you fly into the air a bit and then settle down in a sort of SHM like motion. Looks quite funky. So the physics is properly 3d, but the collision detection is only 2d. I think that this'll be OK though, because you'll always be a pretty much fixed distance from the ground, with minor variations.

How long do you think your project is going to take? It looks like it could be quite a long time. I'm hoping to get a playable version of my game working by the end of the summer, albeit without any AI or complex features. Maybe have it actually finished a year after that.

Probably ok....dunno, you'll find something to mess you up somewhere. You could probably fix it reasonably easily by doing an extra check after the 2d collision happens to see if theyre at the right height. And when 2 ships collide, you just *have* to get the bounce away depending on the vertical positions as well. I.e. when you dive bomb off a cliff, youve got to drive them into the ground on the ram

Length of project? No idea whatsoever. Galileo is just that bit too ambitious probably, and theres been an awful lot of design work gone on, many many people working/worked on it. Well, lets just say that strictly speaking it started in 1996...

Hmmm, you're right about the sphere thing. B****r, I'll have to rethink that a bit. Maybe I'll model it as spheres and vertically extruded 2d polys. I'm trying to avoid doing a full blown 3d thing here, because that's more work than I think I can do on my own.

Maybe Galileo is falling into the trap of always designing and never actually getting round to doing the coding. It's quite an easy trap to fall into, because designing is so much more fun than coding I spent ages doing design for my game, but then I basically scrapped it all, I'd been far too ambitious.

Spheres and prisms shouldnt be too hard. For the prisms just ignore the z dimension entirely. And sphere to sphere collision is easy to test for.

We are well into the coding stage now. Most of the core design work (what stuff you'll actually be able to do in the game) is sorted out, and we've barely started on deciding on the details/naming things etc.

Theres a lot of stuff there in the design, but everything there, at least one of us knows how to go about doing.

Yes, that's what I'll do. Infinite prisms and circles. Shouldn't require too much modification.

How many people are working on Galileo? I'm guessing quite a lot?

It varies, but we had about 12 people on irc at the last meeting.

Just in case you're interested, I finally found some time to do some work on my game and there's now a playable (er, sort of) demo on my website

Hmm, I seem to be missing all3928.dll - any idea what that is?

Thanks for pointing that out, I'll upload it!