At 10/12/11 03:49 PM, turtleco wrote: wat?
wat
HO HO HOPE you become a Newgrounds Supporter this year!
We're working hard to give you the best site possible, but we have bills to pay and community support is vital to keep things going and growing. Thank you for considering!
At 10/12/11 11:59 AM, Sandremss128 wrote:At 10/12/11 11:52 AM, andrease wrote: TLDR; Need to animate shit - fresh out of ideas. Wat do?Maybe something with chewing-gum...
nah
Shit - I might just fucking do that. Thanks.
Fuck the corporate world!
At 10/12/11 06:22 PM, Sam wrote:At 10/12/11 03:49 PM, turtleco wrote: wat?wat
wat
Slint approves of me! | "This is Newgrounds.com, not Disney.com" - WadeFulp
"Sit look rub panda" - Alan Davies
At 10/13/11 01:25 PM, citricsquid wrote: http://lights.elliegoulding.com/
now THAT is an experience!
From the first second I saw it I knew it was to beautiful to be flash. Really nice to see such high quality stuff coming available for internet browsers.
At 10/13/11 08:31 AM, andrease wrote: Shit - I might just fucking do that. Thanks.
To that I drink an ice-cold beer. =)
What is everybody working on? I am working on an easy pathfinding API. I am tired of using other peoples code for such a common problem. I am also playing with ND2D, which I find better than starling so far. I finally forked over the money for a private account on github because I can't be bothered to set git up the hard way. Also I am addicted to triple town.
At 10/14/11 11:44 AM, PrettyMuchBryce wrote: What is everybody working on? I am working on an easy pathfinding API. I am tired of using other peoples code for such a common problem. I am also playing with ND2D, which I find better than starling so far.
I started learning Flixel this week with the hope that it will adapt well to hardware acceleration. I started looking at ND2D as well, are there any good tutorials for people who really need a hand-holding through getting started?
At 10/14/11 11:51 AM, TomFulp wrote:At 10/14/11 11:44 AM, PrettyMuchBryce wrote: What is everybody working on? I am working on an easy pathfinding API. I am tired of using other peoples code for such a common problem. I am also playing with ND2D, which I find better than starling so far.I started learning Flixel this week with the hope that it will adapt well to hardware acceleration. I started looking at ND2D as well, are there any good tutorials for people who really need a hand-holding through getting started?
Well...I guess I'm on the same side as well, I'm somewhere near the intermediate and the beginner's of AS2 and the novice's of AS3.
And now the virus my computer got infected with early this year cored out the Java thingy for AS3, and I can't use AS3 since it wasn't present. Well, I guess my experience with AS3 was over before I even started :\
A clean reinstall may fix the problem.
Slint approves of me! | "This is Newgrounds.com, not Disney.com" - WadeFulp
"Sit look rub panda" - Alan Davies
What makes N2D2 better than Starling? I checked out Star but then got a crapton of schoolwork, so I haven't looked into N2 yet or even experimented with Star.
Fall break hooray!!!
At 10/14/11 11:51 AM, TomFulp wrote:At 10/14/11 11:44 AM, PrettyMuchBryce wrote: What is everybody working on? I am working on an easy pathfinding API. I am tired of using other peoples code for such a common problem. I am also playing with ND2D, which I find better than starling so far.I started learning Flixel this week with the hope that it will adapt well to hardware acceleration. I started looking at ND2D as well, are there any good tutorials for people who really need a hand-holding through getting started?
I'm not sure if there is yet since it's so new. It's pretty similar to Starling in the way you start it up. Maybe I could write a basic "Getting Started" tutorial over the weekend.
I'm not sure if Flixel will transition well to hardware acceleration since its still being rendered with software (BitmapData.copyPixels). I am sure there will be new engines that take advantage of the GPU stuff sooner or later though. Starling and ND2D are not really like Flixel. Flixel is unique because it geared 100% for making games. So it includes graphics, sound, physics, state machine, etc. Starling and ND2D are just for graphics, so you have to write the rest of the game logic yourself. Learn Flixel anyways though! The concepts will carry over.
As far as the difference between Starling and ND2D. I have no statistics to back this up, but I find that ND2D is just faster. It has a nice support for animating your sprite sheets similary to the way Flixel does it. Starling has no documentation other than a massive messy PDF file. I worry that Starling is trying so hard to imitate the traditional display list that it is making concessions with performance. One downside to ND2D however, it has bad support for smoothing. For example, if you're looking to make a pixelly game, you might want to go with Starling for now. If you resize a texture in ND2D then it will automatically smooth it and there is no way to turn this off. The work around I found is to resize your sprites with bitmapData.draw before you make it into a Texture2D in ND2D. Lars says that he is working on fixing this, though. In Starling there are a number of resize modes you can choose from which is pretty nice. ND2D also has a brutally awesome virtual camera. Honestly though it probably doesn't matter too much which one you use. They will probably both change a lot over the coming months.
At 10/14/11 05:20 PM, citricsquid wrote: even if you hate dubstep (fuck you) this animation is sweet:
http://www.youtube.com/watch?v=DbCYsEMg0 9o
whoahwhwh wtf they are doing flash animations for their sponsored songs now sweet! i've had a couple ideas for a few different tracks in my head. can't wait to see more
the events are merely fictional, written, directed, and acted out by all who create them
At 10/14/11 05:20 PM, citricsquid wrote: even if you hate dubstep (fuck you) this animation is sweet:
http://www.youtube.com/watch?v=DbCYsEMg0 9o
this one is sweet as well:
http://www.youtube.com/watch?v=2cXDgFwE1 3g&feature=player_embedded
At 10/14/11 05:20 PM, citricsquid wrote: even if you hate dubstep (fuck you) this animation is sweet:
http://www.youtube.com/watch?v=DbCYsEMg0 9o
There was a cat playing bongos...
At 10/14/11 05:42 PM, Sandremss128 wrote:At 10/14/11 05:20 PM, citricsquid wrote: even if you hate dubstep (fuck you) this animation is sweet:this one is sweet as well:
http://www.youtube.com/watch?v=DbCYsEMg0 9o
http://www.youtube.com/watch?v=2cXDgFwE1 3g&feature=player_embedded
that's.....not an animation
the events are merely fictional, written, directed, and acted out by all who create them
I finished this: https://github.com/prettymuchbryce/EasyS tarAS3
Any suggestions or questions are welcome :)
At 10/15/11 09:52 PM, PrettyMuchBryce wrote: I finished this: https://github.com/prettymuchbryce/EasyS tarAS3
Any suggestions or questions are welcome :)
*Wow, dyslexia on ND2D...*
Since you're using a grid, you don't REALLY need the whole node thing. I wrote a grid-based A* and converted it to draggable node-based later on. Your code is similar to my draggable node one. Since each tile is a tile's width from the next in either direction, it would be better to avoid Point() and try a more grid-based approach like (I haven't seen the grid code in a while, might not be 100% correct)
//it would also be better to switch the vector to booleans, I feel; here true = no obstacle
for(i=-1;i<2;i++){
for(j=-1;j<2;j++){
if(i==0&&j==0) continue; //avoids checking self
if(i!==0||j!==0) continue; //this line restricts it to 4 directions, omit for diags
try(gridVec[currentI+i][currentJ+j]){ //code that adds it to the potential node checking thing}
catch(e:Error){}
/*and so on...I guess currentI and J could be like int(currentTile.x/tileWidth) or something*/
}
}
And if you do allow diagonals, you would have to change your heuristic to x^2+y^2 rather than x+y
The grid just stores the collision data. 0 means you can walk on it and 1 means you cant. As far as the Node class, I think it's the best approach since I need to store values associates with each node (G, H, and coordinate), right? :)
At 10/16/11 03:12 PM, PrettyMuchBryce wrote: The grid just stores the collision data. 0 means you can walk on it and 1 means you cant. As far as the Node class, I think it's the best approach since I need to store values associates with each node (G, H, and coordinate), right? :)
If it's only gonna have 0 and 1 as a value, might as well switch to booleans to speed it up a bit. The generic node class is fine (I use that), I meant using Point() in the algorithm when it can be shortened and optimized by using its .x and .y to find its location within the Vector (vec[i][j]). I'm not fond of Point(), and you will get better performance avoiding it. My grid one clocked the algorithm alone at 5-8ms on a 20x20 grid with a normal path, 25 with a complex one, and 53 for no solution.
I'm kinda happy your code looks just like mine...I guess we're both doing something right :)
At 10/16/11 04:29 PM, MSGhero wrote: I'm not fond of Point(), and you will get better performance avoiding it.
Okay you are right about that. I changed it. Thanks.
Speaking of Flixel I stumbled across this cool article today about getting started with Flixel Development: http://www.photonstorm.com/archives/2247 /flash-game-dev-tip-12-building-a-retro-
platform-game-in-flixel-part-1
Anyone have experience with Flixel? What did you think? I'd be interested to know. I have used it and I learned a lot from it. If you're interested in fast development and don't care too much about fine-tuning stuff then I think Flixel is a good solution. If you spend enough time with it to go deep into the flixel classes you will learn a lot about how flash works and how to make a flash game perform well.
And here's my A* prototype that I've been working on (with a little help by PSvils). Running my A* algorithm through an irregular city each frame at a cool 30 fps. It's also my first project that I'm using external classes and a timer-based enterframe :). The blue dot isn't moving on a set path or anything, so it might overlap the obstacles at some point.
2 things: why does the mem usage go up dramatically and ms per frame freak out testing in browser than in Flash Pro? And if the blue dot could move along paths and stuff, how could I tell the other nodes that it's visible for them (quadtree, etc)? It can't be visible all the time, cuz then it would just draw a straight line through obstacles.
The memory shows the total memory used by Flash in the browser.
There are much simpler ways to do pathfinding in that sort of setting. I've talked about it many times. I might put up simple a demo on Github.
What is your other question?
At 10/19/11 10:14 PM, MSGhero wrote: And here's my A* prototype that I've been working on
I got a question. At some point you have the line go from the main node to eventually node A. From this node A it goes to the moving circle. But in some positions, the path still goes to node A but in a different way. Why is this since these nodes dont change position and the shortest path should be fixed for them.
At 10/20/11 04:59 AM, Sandremss128 wrote: I got a question.
Thanks, the x and y positions were ints in my grid-based version, and I forgot to switch them over to num. Works fine now, I guess it was rounding or something weird.
My other question was there, you might have missed it. About walking down a path and quadtree visible nodes and stuff. How else do you suggest pathfinding? To be honest, it will only need to be run maybe once every few frames to few seconds, depending on the eventual map size; so it's not THAT big of a deal. Check where it can go, check how far it's been and will have been, see if that's smaller than another's.
If you're talking about the red dots or the lines, those are just visuals for me; it stores the dots' positions at the start; they aren't referenced or anything later.
If the positions of the nodes (dots) dont change you might even consider making a table for it upon initialization. That means the CPU is a little more busy loading your level / game but when you play it it executes faster. And I dont quite get your question, you just do the A* (A star) algorithm and then you've got your fastest route? (btw A* is invented by Dijkstra, a Dutch person :D)
I mean rephrase it. Visible from them?
http://www.newgrounds.com/dump/item/05ca 72840c163be12766c70fc92c3263
A* is not the same as Dijkstra's algorithm. A* extends the latter with cost heuristics, which allows depth-first searching.
The more you know.
I know but what I mean is that the foundation of pathfinding comes from this guy.
Um, he established all of computer science. If you want to feel proud about something.
But he defined a special case of a toy algorithm too.
really fun discussion to read, thanks for making the reg lounge an interesting place for these last couple of days
i've been away from computer science way too long, forgot how fun it is
