Newgrounds.com — Everything, By Everyone.

At 2/3/09 03:20 PM, GustTheASGuy wrote: Angle to Cartesian is sufficient to get coordinates. Just add up the angles.
I can't see how the robot would be complex enough to be able to break itself if you just limit angles of the joints.

Wouldn't the twisting throw it off? The third joint changes the plane in which the next two joints work...

idk.

I'm working with java3d right now. The only think I'm worried about is if I can get absolute coordinates for objects. I'm sure you can, I just haven't come across it yet.

At 2/3/09 01:14 PM, GustTheASGuy wrote: I'm pretty sure the 3d API you're talking about is for drawing graphics.

yeah, but I was hoping for two things
1) that it handles the transformations for me (which I'm sure it does)
2) that it can show a 3d model. Not necessary, but it would be nice to make sure it's all working right internally.
of course, I need to be able to detect collision too, one way or another.

Do the joints have different planes of rotation?

yes, the first one only moves parallel to the ground, the next moves up and down, the third twists, like a wrist, the forth: god knows what it is because of the twisting of the third.

the forth and fifth i think both move in the same direction, which is up and down when the third is in its default position. The sixth is another twister, i guess, and the seventh is a gripper.

if it wasn't for all the twisting, I think I could have figured all of this out with my existing knowledge. Probably a spherical-to-cartesian coordinate conversion could have solved it.

So then I started thinking about transformation matrices and stuff, but I haven't had linear algebra yet, and the stuff I look up online doesn't make any sense. That's why I figured a 3D graphics library would be good, because isn't all that stuff already implemented?

The most elegant solution would be to distribute motion to all joints and simply make sure they don't bend too far. You could of course model (not in the graphical sense) the 'bones' (segments) and ensure they don't intersect but that might lead to the robot being stuck if it bends wrong.

I'm not entirely sure if I follow. You can't exactly limit the angles of specific joints, because they all depend on the other joints before them. You could set up cases, which we did, but it's not to long before you realize there are too many.

is that what you're talking about?

ultimately, I need to know the positions of every joint in a 3d space, not just to keep it from hitting itself, but we're eventually going to implement a algorithm(GA maybe) to move the arm, and so we need to be able to calculate the arms exact position with the angles of each joint.

I'm working on a robot at school and it has an arm with 7 joints. the problem is, the robot can easily damage itself (because there's no restriction in its movement). So it's my task to put 'bounds' on the arm so it can't try to move the arm through itself.

I thought the best/easiest way would be to model it's arm in 3d, and check for each move, for each joint, if it is within a certain bounding box around the robot.

We're using java, so I thought of using the Java 3D api, but that's something extra, and im not even sure if that's what i need.

Each joint is obviously dependent on the joint before it. each joint only takes one angle as a parameter.

Back when I messed with 3d graphics I remember using bones, I think if I had something like that, it would be good... But I didn't see anything about bones in the Java3D, though I didn't look too hard.

Another option is to work out the 3d transformations without a 3D api. If I had to do them myself though, I couldn't be sure that it was actually working right, and there's not really any room for error, we already broke the arm once...

So if anyone has any ideas or suggestions I could really use them.
Has anyone here ever used Java3D API?

You mean XP doesn't have built in tools to shrink partitions?

Vista can... I guess that's one thing they got right...

At 12/15/08 01:16 AM, henke37 wrote: Data structures, needs more data structures.

yeah, you need an AVL tree.
Self-balancing trees are the shit.

I was going to contact the website and tell them they need to get their browsers straight, but when I scrolled to the bottom I found an AOL email address.

I figured I would be wasting my time.

At 10/28/08 10:01 AM, DearonElensar wrote: So why are we bitching on people using Game Maker then?

It may still be considered programming, but the usefulness of learning and writing Game Maker games may be minimal. I think it is good to encourage to use Game Maker (or ALICE or whatever) to learn and/or make quick and easy programs/games, then to move on ASAP to something that is more flexible and useful.

But if someone really just wants to make a game as a hobby, and doesn't really want to know any other aspects of programming, then you shouldn't tear them down for using Game Maker. It's fine for what the want to do.

At 10/28/08 11:21 AM, authorblues wrote: (mapping a space is not a linear regression problem)

I didn't really think so

:if youre just wanting to filter out noise , a kalman filter should do nicely.
The kalman filter definitely looks promising. I'm going to look at it.

The points that I'm getting can be automatically converted to cartesian coordinates, and that's the info that I'm storing, the x and y of the reading. Erroneous readings can be scans that get reflected and never return, producing a far off reading (seems to be capped at 5000mm though), can also see obvious error at greater distances and at angles not perpendicular to the wall (though this is to be expected).

The sensors do not rotate, they are fixed to the robot.

I guess what I really want to do, is be able to tell the difference between static and mobile elements around the robot. I would like to keep track of the position of the static and mobile elements, and also the velocity of the mobile elements. For example, one 'demo' I want to do is have the robot follow a 'leader'. Right now it works pretty well, until the robot decides to "follow" a wall instead of you.

At 10/28/08 12:28 AM, VigilanteNighthawk wrote: First, let me put out the disclaimer that I'm hardly an expert on AI. I've only done some light reading on it.

I don't think this sounds like a line of best fit problem. I could be wrong, but it sounds to me like the task is to get the robot to create a "map" which consists of straight line paths to points of interest in the room while filtering out noise, which would be your random points. If this is the case, my guess would be that you could use a neural network to train the robot to filter the noise. For there, the mapping should be relatively straightforward.

Yes, I do have to do that. But the thing is, The readings from the sonars can be in front of, or behind the wall being sensed. And what i want to do is take all those points and figure out (to the best of my ability) where the walls are. So that's where the line of best fit comes in... Obviously though I have to trim that line into a segment, but that's not difficult.

It is important that i figure out the walls from the points.

1) I never said it was

2) I'm not implying that the future of programming is going to be like ALICE. It is slower to use ALICE. But I don't see why we won't possibly see more Visual IDE's.

IDEs are becoming more visual. And there is an interest in Visual Programming, dragging and dropping programming elements in a flowchart can be easier for a lot of simple programs. And especially with the focus on modular programming, I believe it can and will work really well.

And also, the speed at which you can enter code is not the only factor to consider when picking a language/ programming environment. Sometimes if it takes 10% longer to enter it in the first time, but is easier for others to modify or for you to debug, it can be worth it. And that's where I think Visual IDEs will win out. ESPECIALLY with businesses. (think of the ridiculousness of COBOL and tell me that anything is impossible...)

Does anyone know of any solutions to the following problem:

Given a set of points in a 2D space, find the line or lines which represent the scattered points the best. Knowing that some points can simple be random points which should not be represented...

In my AI class we're working with a robot, and the points are sonar readings. I'm trying to figure out the best way to map the room with these readings.

I've been thinking of "the line of best fit" in statistics. This would kind of work, but there are many lines, and how do I know which subset of points to make a line out of?

So has anyone ever heard of anything that would be good for this?

That's funny that you guys say that. I watched a video online where microsoft was unveiling it's new language F#, and the guy talked about 30 years ago developers imagined that we WOULD be programming in drag and drop environments. Surely after 30 years we wouldn't still be typing instructions into the computer! absurd!

Who are you to qualify what is programming and what is not? What gives you the right? How about anyone who doesn't code their programs in a HEX editor is not really programming?

The fact is, programming does not HAVE to be what it is now, and arguing that point is extremely naive. Development in programming IDE's have not really improved much. So really, just because we're still in the stone age of IDEs, doesn't mean there's no room for improvement.

Drag and Drop programming can equal programming, especially since Alice is Java at heart, what's the difference? if I copy and paste did I code a project? what about code reusability? This is essentially just that...

Also, I don't know much about game maker, but I would imagine it could be considered programming as well.

A quick Google Search returns this definition for programming:

the act of creating software or some other set of instructions for a computer.
cyber.law.harvard.edu/readinessguide/glossary.html

The act of creating some set of instructions for a computer. Whether that be compiled machine code or interpreted byte-code or an constructing the rules for an expert system, or creating a bash script or a batch file or making a javascript page or making a bloody freaking game maker game. According to Harvard's definition, those are all programming.

There is no qualification based on HOW those instructions are given to the computer. In the end, it doesn't matter.

ALICE is good at what ALICE does. Game Maker has it's purpose. C is good for getting down and dirty, but not everything needs C, or is even a smart choice for C. Not everyone should learn on C, or Java, or whatever. The reasoning behind ALICE is to let the student focus on the logic of programming rather than the syntax of a language. Believe it or not, there is a rather large learning curve to programming, and ALICE smooths it out a little.

And java is not crap. That statement is simply dumb. Java is crap at X, or Java is good at Y makes sense. And VB is a programming language and is a lot better than most "skilled" programmers will give it credit. Believe it or not, not everything is tailored for your needs.

So yeah, recap:
-Drag and Drop programming was at one time considered to be the future of programming.
-Java is not crap. Java is good at what Java is good at. It may be crap at certain tasks, but that does not make the whole language crap.
-(Drag and Drop != programming) == false; // The method of entering instructions into a computer is irrelevant.
-Therefore Game Maker and most certainly VB can be considered programming.

I personally don't use ALICE, but at my university we are seeing less dropouts of computer science related fields now that we're using half a semester of ALICE before getting into Java. Whether or not that's simply because it's easier, well, we shall see. Theoretically, it's a good idea, and students should understand more.

At 6/14/08 05:12 PM, yhar wrote: Well, i was told that you and tina had some sort of agreement based upon your claim to sales, then it turned out all of your customers were "kids" with no intent on renewing their hosting.

Yeah, noone had any intention of renewing their hosting because that bitch fucked everything up. When the switch was made, I could not log in to the axishost site, and i could log in to the adm site, but none of the controls there worked. I kept making changes using cpanel, and sending her e-mail after e-mail about not being able to log into the site. I also stopped receiving the automatic payments on paypal... I wanted to keep my service, I really did.

Long story short, she never got any money from me, but that's her own damn fault. She also ripped me out of a domain name. goDaddy said that I had to wait till after it expired, but of course, as soon as it expired, someone snatched it up.

var k = 1;

while (k > 0) {
    trace ("I will not make stupid forum topics");
    k++;
}

AS3. Is there any way to tell when a unit is no longer in a movieclip?

For example, I have a unit that gets created at a certain frame, then later on, it disappears. I have an class that corresponds to this MC, and the constructor adds the unit to an array. Because of this array, the MC stays in existence longer than it should (you can't see it, but it keeps shooting from offscreen causing lag.

I tried testing to see if the MC is a child of the parent MC, it still is, and I also checked to see if it's still visible, it is. So my question is there some kind of Event or something that happens when a MC "erases" a DisplayObject because it is no longer needed on the timeline? Or is there some way of accomplish what I want?

Thanks.

I'm willing to help you out. Check my submission Desert Outpost Defense for recent work (in AS3).

At 6/10/08 08:51 PM, starBlinky wrote: thanks guys. hmm. I will try and continue. But I'll probably shorten the level amount to like 10.. That's how much I hate making this. lol.

you should make like, 20. Do that, and then make the graphics a lot better, add some music and sound effects, and I bet you could get a really good sponsorship for it... It's a very interesting puzzle game.

it's good, please continue!!

though it is hard sometimes to hit those blocks at full speed.

At 6/10/08 03:43 PM, nns357 wrote: Congrats on being the most ignorant person of 2009, already!

Shutup. seriously. You're making a fool of yourself.

Obviously the guy has no intention of doing this on his own, he wants to pay someone to do it (in which case he should probably try rentacoder.com or something). So let him do what he wants; it isn't better for everyone to do this stuff on their own. Would you expect a CEO of a company to do ALL of the minor tasks within the company? no, he tells other people to do it (not to say this guy is a CEO or anything).

Obviously this web design stuff isn't this guys cup of tea, and obviously he has enough money to justify having someone else to do it.

Check the placement of your functions, are you sure they're where they should be? And not within something else?

That's my best guess.

There is an easier way, use the Array's join() method. It joins all the elements together, but separates them with whatever you pass to join. So if you pass nothing to join, it shoves nothing between the array elements.

You should become familiar with adobe's liveDocs. A quick lookup of the Array class would have told you all you need to know.

Once you compile your flash movie, the layers you have at design time no longer exist. Flash looks at the layers, and assigns every object on them a depth based on what layers they are. So at runtime, you have no layer2.

The easiest solution would be to add a blank mc to layer2, and then add whatever you want to the movieclip located on layer2.

The hard solution would be to shift all the depths around manually, or decide to add everything at runtime.

-the collision detection, its a little odd. Sometimes I get hit when I am clearly not in contact with the enemy. It seems most glitchy when I'm moving. I actually tested where if I was standing still and an enemy passed going down, I could not get hit, but when I hit down and chased after it, even though I didn't change my horizontal position at all, when I passed it and was ahead of it, it killed me.
-Different types of enemies.
-The only pickup I ever seemed to get was "slow".

I was planning on doing a top-down view, since that would be easiest to draw. You would have to come up with different components for the bots, mostly weapons, probably. Different Tilesets to build levels out of would be your responsibility. Maybe different types of effects for bullets, lasers, etc... Also, designing a menu screen and the user interface. It probably won't be too difficult, but I suck at the art stuff.

It wouldn't be just art, that's why i said I need design help to. I would like an individual who could come up with ideas for all this stuff on his or her own. I'm looking for someone who would come up with ideas about the game, too.

At 6/4/08 02:06 PM, UknownXL wrote: Show some proof that you are capable of finishing this game, and maybe some more examples. No one wants to start something that wont be finished.

All my examples are at timegames.net, and they also all happen to be on my user page. I am a computer science major, and fully capable of finishing a game.

Not only that, but who cares if it doesn't get finished? I said I would pay, so you benefit whether or not I finish. I realize now that I said "opportunities", but I meant that you will be paid. Oh, and payments are via paypal or not at all.

I am working on making a game that will be similar to Bot Arena [gamegecko.com]. I need somebody to be able to do the graphics and perhaps also help out with design ideas. The graphics need not be in flash, just high res, since I'm probably going to handle everything as bmps anyways.

This project will probably be slow paced from your perspective, since I am working on some stuff that is rather difficult programming wise (I am NOT just making a clone of Bot Arena I have a very distinct goal in mind which is different that most any game out there). So there would be no real pressure on your part, but I would expect regular contributions on your part early on, so I have stuff to work with.

Also, I will pay, though payments would have to be worked out with the individual.

If you really liked Bot Arena, and have some good idea's to improve the game, and can draw well, please post back here with links to previous work (or example drawings or something). You can also PM me to discuss payment, but just don't pm me so the thread remains active.

Oh, and for examples of my previous works, visit timegames.net.

I am working on making a game that will be similar to Bot Arena [gamegecko.com]. I need somebody to be able to do the graphics and perhaps also help out with design ideas.

This project will probably be slow paced from your perspective, since I am working on some stuff that is rather difficult programming wise (I am NOT just making a clone of Bot Arena I have a very distinct goal in mind which is different that most any game out there). So there would be no real pressure on your part, but I would expect regular contributions on your part early on, so I have stuff to work with.

Also, there is opportunity to pay, but not much (I'm on a low budget). Payments would have to be worked out with the individual.

If you really liked Bot Arena, and have some good idea's to improve the game, and can draw well, please post back here with links to previous work (or example drawings or something). You can also PM me to discuss payment, but just don't pm me so the thread remains active.

Oh, and for examples of my previous works, visit timegames.net.

I hope you've learned to back up your stuff after this.

Almost every day I save a backup copy, for the different versions. Roughly twice a week i save them to a flash drive. Every once in a while I save it all to a completely different flashdrive which I leave sittting in my room.

For comparison, Google File System (GFS) stores three copies of its data automatically in three different locations, it also handles routinely comparing that data and fixing problems with corrupted or lost data.

For anything you feel to be important, you should do no less.

And ESPECIALLY when you're going to change something, either hit Save As..., or just do what i do and zip the folder it's in and rename it as "XXXXX backup vX.X.X"

At 4/22/08 10:24 PM, flashash220 wrote: Does it matter? linux has no warrantee
No safe distributes
It's open source
Overall... it sucks
Leave the Os to the pro's at M$

Wrong?

Does it matter? linux has no warrantee

Doesn't matter since you can fix it yourself when it breaks.

No safe distributes

No idea what you mean.

It's open source

I thought you were arguing against Linux?

Overall... it sucks

The only thing that sucks about it is compatibility problems. The Linux community tries very hard to make everything compatible with Linux, but if software and hardware vendors would take the time to port to *nix then we wouldn't have that problem. That will change (and is changing) over time as Linux becomes more popular.

Leave the Os to the pro's at M$

Benefits of Microsoft:
-Pay to do everything
-When it breaks, hope M$ can fix it, or you're SOL
-Format Lock-in
-Minimal control over your OS

Just to name a few.

Now I'm not trying to start an OS religious war, but I can't stand it when people blindly eat up Microsoft's crap, and try to tell people that FOSS sucks because you don't pay for it.

Linux is a good alternative. Especially for servers.

I finished a game not to long a go, and I need a way to spread it around the internet. Is there any easy way to get it to as many sites as possible?