This is very slow-paced, I wonder if the gameplay could be made more hectic somehow without impacting the (excellent) feeling the music/graphics give. At the moment I feel a game goes on too long.

I think you could do with smaller levels. There's a lot of empty space which is there to make you walk around rather than because the space is required for a puzzle. Walking is not fun: try making each level as small as possible (ref: level design workshop exercise 1). This problem was particularly pronounced in the introductory levels: every outside level could have been a single non-scrolling screen without losing anything.

I think you should add new gameplay elements more slowly and focus more on the puzzles. There aren't any actually interesting puzzles until level 11 or so, before that it is all simplest-possible-puzzle for each new item. That is not fun in of itself!

Movement is slow and feels a bit unresponsive. I think it's possible for key-presses to be lost if you're moving when it's pressed?

"You are an adventurer": don't say this, it's obvious anyway but it also breaks immersion and just isn't good writing (i.e. show don't tell).

Level 10: blocks on top of fire should have some visible indicator that they're on top of fire! I agree this level is too long. All of Increpare's comments on this level are spot-on, it's too long and there's too many places where you can die after encountering a new mechanic: forcing you to replay the whole level again. The message "Lava requires both fire and water protection" has never been so annoying...

The behaviour is unexpected when you move onto ice without there being normal floor across from you. I expected to be stopped by the wall and then to be able to choose a new direction to walk in, instead I bounce back. Not really a problem I guess, just something I thought was a bit weird. I also find it weird that a wood block pushed on top of water doesn't react (I'd expect it to create a walkable platform).

I am not sure how much some of the enemies add to the game but maybe that's just me being a puzzle purist. Since it is a tile-based game, avoiding enemies early on is just been about timing and isn't particularly interesting or skillful (note that "oh crap I have a very short time to press this button or I'll die and have to restart the whole level" does not require skill, it just creates frustration). Some enemies like that rat have puzzle implications that work fine so possibly it is just a few enemies that aren't necessary.

Overall conclusion: it's certainly a promising set of building blocks for puzzles, as I said before I think the puzzles themselves could be streamlined though.

Sound-delay: Linux has an unavoidable half-second delay on every sound. If Increpare is having issues not on Linux not sure why that might be: that code snippet looks fine.

I think I remember hearing something like "each time you play a sound for the first time in a scene (or something) flash has to start up the audio subsystem".

I think this was a problem with older Flash players and shouldn't be an issue with AS3. Could be wrong though.

Why did the Englishman only play Chess in Prague?

Because he had a Czech mate!

You want to use Event.ACTIVATE not FocusEvent.FOCUS_IN.

Not sure I like that much from a user perspective but if that's easiest for you then okay.

Would you notice the reloading if nothing has changed? Wouldn't it just be replacing the history with an identical history? You could also just check for new items in the SharedObject that you don't have locally.

What I would initially test would be always flushing after writing, and re-reading the SharedObject when the app receives a focus-gain event.

If that doesn't work I would probably investigate inter-SWF communication with LocalConnection.

Your sign-up page does not specify anywhere that it wants an email address. That may be obvious but it is not great UI to have a textfield with no explanation.

What will you do better than any other metric-tracking systems, for example Playtomic? Note that the usefulness of an application like this is heavily dependent on the interface you provide to view and interpret the data; simply logging events is something that anyone could set up themselves.

Filters > Map > Map Object > Map to: Sphere

Doesn't seem to have a torus option.

[I would really appreciate feedback on this change]

New version!

Changed the jump mechanic completely, now there's no more charge jump, jump still beats block but attack beats jump. I would really appreciate feedback on this change, especially two-player feedback if you can grab a friend.

AI difficulty noted: this has a new AI which is hopefully better? Though probably still quite hard... The old version still has the old AI.

You can now fight the AI yourself: I would appreciate feedback on the difficulty.

Working on some AI:

Yes, we already have those, they are called compilers C++ is too painful to parse for this purpose without basically a full compiler infrastructure, imho.

Painful to parse properly, but it should be easy enough to hack together something inflexible (and probably incredibly brittle) that would work for a small project.

Not that I'm really aware of, however I've found when someone starts charging the jump when I'm not expecting it, I often don't have time to react. So it's a risky strategy but it can sometimes pay off even at short ranges.

I am not opposed to replacing the jump behaviour but currently don't have any alternatives in mind. It needs to have a long charge-up time because it gets past blocks.

Glad you had fun!

Hey guys Im in the middle of completing my verses entry. However Im finding it difficult to complete the game by myself primarily because I suck with maths  Any flixel/flash programmer willing to help me finish the game please take due credit. No network programming involved, very basic stuff !

Thanks

You probably want to provide some more details on the game if you want to entice somebody to finish it for you. A good idea would be to start a thread describing the game idea and providing screenshots.

Also you should introduce yourself!

[FUCKING AWESOME]

Rotating a bitmap object using the same built-in functions, .draw()'ing it to a bitmapData, then copyPixel'ing that to a canvas gave 44 FPS.  This likely would have been much faster if I'd rendered the rotated bitmap just once then shared that bitmapData between all 100 instances, but that didn't match the situation I was comparing solutions for.

If you're not caching the rotated image, why do you need the intermediate buffer?

Earlier today I implemented a 2D lighting system where light is attenuated exponentially as it passes through substances of different opacity (for now, solid and air) and is then rendered in discrete colors.

It's FUCKING AWESOME and it was only about an hour and a half of work including tweaks.

Sounds cool, do you have a screenshot?

I think the gain is chiefly in cutting out Flash's display list. The you get secondary gains for not using scaling or rotation (translation was already in my test).

Hmm, It was my understanding that the primary benefit was from cutting out the vector renderer, and the display list overhead was a significantly smaller factor.

I am assuming your DisplayObjects have mouseChildren = false and mouseEnabled = false?

I do it like this:

Code:

private static function test (a:*, b:*, c:Class):* {
if (a is c) return a;
if (b is c) return b;
return null;
}

public function doSomething (a:*, b:*):void
{
var bullet:Bullet = test(a, b, Bullet) as Bullet;
var enemy:Enemy = test(a, b, Enemy) as Enemy;
var player:Player = test(a, b, Player) as Player;

if (bullet && enemy) {
// ...
}

if (bullet && player) {
// ...
}

// If I want to detect bullet-bullet collisions...
var bullet2:Bullet = (a is Bullet && b is Bullet) ? b as Bullet : null;

if (bullet2) {
// ...
}
}

Iain Lobb has also done some similar comparisons.

Err - why wouldn't you just use copyPixels? Using draw() to copy a BitmapData when there's a much faster method is a bit nuts.

It's relevant if you want to render a BitmapData with scaling/rotation.

I keep meaning to test BitmapData vs. display list performance when you want scaling/rotation but I've never got around to it...

Updated again. People were having issues with keys interfering with each other when one player charged up a jump so I've moved to a three-button system.

Don't think I like this new control scheme much, but can't see much of an alternative.