Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Muz]

I do wonder how companies like Oracle make money.

[Siquo]

Consultancy. I've worked with Oracle products as a financial analyst for a multi-billion-dollar company. That thing was a monstrosity...
It's a top-down cascade of companies (Oracle is one of em) throwing bags of money at stuff to make it work.

[MagmaMcFry]

throwing bags of money at stuff to make it work.

Sounds nice.

[Skyrunner]

So, like Max suggested, I implemented the flyweight pattern, but ...

ARRGH MEMORYLEAK Dx

Back to the drawing board.

Also, the flyweight saves 1.5 MB of ram, so about 10% off.

[Siquo]

throwing bags of money at stuff to make it work.

Sounds nice.

If you're a consultant able to bullshit non IT-people, this is how you make shitloads of money.

[Max White]

So, like Max suggested, I implemented the flyweight pattern, but ...

ARRGH MEMORYLEAK Dx

Back to the drawing board.

Also, the flyweight saves 1.5 MB of ram, so about 10% off.

Do you know how Creator Pattern works?
Well it also defines what class is responsible for deconstructing objects, in most cases. There are exceptions, but it is the best possible starting point.

[Skyrunner]

I actually do know that, but I didn't know it was called GRASP. I just knew that that's how it works. Dx

[Max White]

GRASP stands for General Responsibility Assignment Software Principles, but there isn't a single wikipedia page for Creator, instead the GRASP page covers it. 
GRASP patterns are a lot less of a structural thing, and a lot more like a general over arching guideline.

[dreadmullet]

In a 4 dimensional vector, the components could be called x, y, z, w. In a 5 dimensional vector, what comes after w? What about higher dimensional vectors? Would you just scrap the letters and use the [] operator?

[MorleyDev]

Not sure if it's a "standard" or not but I've seen and used x y z w h d before. Not sure what you'd use for a 7 to 9d vector, i j k?

[Siquo]

w,h,d, is for width, height, depth, for extra dimensions u,v,w are more appropriate. Personally I reserve i,j,k for iterators.

[Siquo]

Doublepost because different subject:

Making a player-object move smoothly in a physics-based platformer is pretty hard. So far I have:
Jumping is an impulse upwards when on the ground. Holding the jump button applies a continuous upwards force while your vertical speed is still upwards.

Moving left or right is applying a force if your current speed is below the maximum speed, and there's a separate force for being in the air and on the ground. Perhaps a different maximum x-speed is warranted when not on the ground.

Now I also need braking, and use friction with the forward force so you can't accelerate as fast when on very smooth surfaces. The controls feel...clunky. Anyone got advice on this?

[Max White]

At some point you are going to have to have a groundedCheck() to tell when you can jump, or at least handle something similar in the collision detection. At that point you should check for a drag value for what ever you landed on. If you have a tile based world, be sure to set it as a new value instead of just taking it out of the actors velocity, so that if they are standing on two tiles they don't suddenly move at a crawl for no reason.
That should make your moment in the air feel more airborne than just running, will cause you to stop when you stop holding down the move keys, and will let you use really low values for ice to slide on, or really high values for mud.

[Siquo]

Oh, forgot to mention, I have that part already The jumping works pretty well.

It's just the walking that's off.
I can use the combined friction (boots_friction*ground_friction) as a modifier for the acceleration force when on the ground. Being used to "perfect" platformers (mario-style), everything that uses physics feels clunky. A sideways-sliding-force doesn't really emulate walking, either, as going up even small slopes (The world is not tile-based but composed of arbitrary shapes, right now just an endless sinus function) is very slow, unless you've built up speed, where it will launch you into the air... Both of those effects are unwanted. Emulating a running person is quite hard when you can only do a wheel or sliding block.

[Max White]

Yea, that isn't the best idea. Anything with a higher than normal traction value designed to stop you quickly will shoot you when you try running on it. Better to split traction and drag into two separate values. Dirt having normal traction and normal drag, cermet having high traction normal drag, ice low traction and low drag, mud low traction high drag, you get the idea. Point is that you don't want to try and force values to do things they conceptually don't do.