You only need physics if you want realistic glancing collisions, etc. (e.g. if you want a stack of crates to fall down properly).

You can still have physical objects. Give them the solid behaviour. They won't move but will still cause collisions with other solids and objects with platform, 8-direction, etc.

What I would do is use no behavior, and do it manually. It is very simple, you do not even have to use vectors!

Instance variable: speed

Every tick: move speed*dt pixels at ship.angle

Every tick: set speed to self.speed * 0.something (if you want deceleration)

Is right arrow down: rotate clockwise -something*dt degrees

Is left arrow down: rotate clockwise something*dt degrees

Is up arrow down: add something*dt to speed.

OK, sure Ashley! This could of course be something to be left to the event sheet.

As I said before, the problem is is that Google Chrome has noticeably better performance in the general aspects of HTML5. I am certainly not going to direct users away from it, especially seeing it is now the #1 web browser in terms of popularity.

I suppose we can do nothing except for getting people to star the issue. Perhaps it would be worth lots of us e-mailing the Chrome representative who has been CC'd the issue (you can see it on the bug report page).

All right, I managed to achieve it with invisible collision objects. However perhaps Ashley could consider adding this as an official feature?

That would work, however the problem lies in finding the angle of the wall relative to the player, since my game is gravity-shifting and rotating.

And still, this doesn't seem to be the best way to solve the problem, I am sure there would be issues...

This isn't in the bugs section because I am not sure it is a bug.

Basically, I have noticed that a platform-enabled object will 'stick' to walls of an incredibly steep slope (anything short of vertical). This is a big problem as most people don't want their players being able to make their way up cliffs.

Here's an example. Totally default behaviors in a new project.

<img src="http://i.imgur.com/tl0mu.jpg" border="0" />

It gets even more insane!

<img src="http://i.imgur.com/FXbcZ.jpg" border="0" />

Is there any way to fix this manually? Or could Scirra implement a solution? Because I am sure this is not what 99% of platform games need!

Hmm. So would you suggest removing my webstorage high score features, and going with Kongregate's API alone for the sake of continuity?

OK, so my question is this:

are Kongregate statistics tied to the player's account (I suppose they are)?

In that case, if a player is on another computer (or even clears their cache!) the stored statistics will be incogruent with the Kongregate statistics!

So if the above is true, then does Kongregate itself notify the player if they have beaten their high score?

Bump - received no reply.

Adding to the topic: If webstorage is used, then a user on a different browser or a different PC would not have their highscores saved: a big problem if Kongregate statistics are tied to login details.

So, I need to know if we can retrieve statistics for a user. If not, does Kongregate's API work well enough that we do not even need highscore feedback in our game? If it alerts the user to a new high score, then I suppose I could just remove the highscore notifications from within the game.

R0J0hound, you have done it again. Many thanks.

Kyatric this would be good for the FAQ I'm sure.