Sprint 3 Demo and Reflection

The investigation has gone better than I hoped. With only a little work I was able to implement a 3d perspective, 3 dimensional gravity and collisions.

I did come across a couple of problems that I’m going to seek advice on.
The Z frame buffer seems to be working in reverse. I’ve got some experience with depth checkers and I’m sure I’ve just got something backwards but I can’t find it.

I can’t figure out how to implement the chase camera. As soon as I have this I can look at implementing a better control system. I’m going to speak with my supervisor about it on Tuesday and go from there.

What I would do differently?
Nothing. I’m happy with the use and result of this sprint.

What did I learn?
A sprint aimed at research and planned experiments can be every but as productive as a sprint dedicated to implantation.

Sprint 3 Planning

Now that I’m getting a grip on this project I’m going to look at expanding into the 3rd dimension.
I’d initially abandoned the idea of doing this project in 3d due to the problems I was having using the framework. Now I’m thinking that I should give it a go.

So sprint 3 is going to be exploratory. I’m going to take a slightly XP approach and write a bunch of experimental throw-away code and see if I can’t get a hold on the Z axis. I’ll implement what I can and identify the things that are going to cause me trouble. Then speak to my supervisor about it all.

Sprint 2 Demo and Reflection

I’m very proud of myself right now.

This sprint started off well but got very rocky pretty quickly.
I set off to correct by collision handling so the rocks weren’t just reversing direction when they collided.
Turns out that the gravity wasn’t working anywhere nearly as well as I’d hoped. Some problem with my implementation made managing collisions impossible.
So once again I found myself working on the gravity forces instead of moving on with the project.

Anyway, I’ve revamped my gravitations, it now less resembles real gravitational forces and more resembles a very simple pursuit model. Basically all of the objects encourage the other objects to chase them at speeds calculated by their mass.
Once that was completed the collision work went relatively smoothly. I say relatively because at the level where the collision responses are being handled I only had access to one of the objects velocity, which made working out the forces pretty challenging. But what I did have was both their masses and positions. So my algorithm assumes that all colliders are moving at least a little bit, it then works out the trajectories of the collision from both object positions and combines, multiplies that by the collidee’s mass and applies the result to the collider’s speed with a 40% reduction to make it feel a little more real, and controlled.

It’s still not perfect but it’s much better than it was, I’m hoping that as I get more practice with my vector math I’ll be able to more finely tune it.

What I would do differently?

Don’t worry so much about getting stuck. Even if you find yourself redoing work, of course it’s not ideal, but it’s still a step toward completion.

What did I learn?

Smaller bites certainly works. I pretty much chose a single thing and went for it. Now here I am having completed what I wanted to right on schedule.