Hey GG, how about a vehicle collision fix??

#1 Here's what I think is Garage Game's mindset on these types of issues:

- Torque was derived from a commercial engine.

- This engine wasn't intended to be used by the public as a generic engine.

- When the public version of the engine was made available (Torque) a lot of code from the old Tribes2 days was left in. Some of this code works well, some not so well and some not at all. A decision was made to leave in as much code as legally possible, despite the fact that it may not perform well. Garage Games consider this a bonus to the customers. Sure, the code may not behave desirably but it's better than nothing.

- The reason code like the vehicle code has been left in the engine is it serves as a starting point. Rather than having to create vehicles from scratch you have a base to work from.

That's a pretty brief explanation and is only speculation on my behalf. I come to this conclusion from things that I have read from Garage Games employees both on and off this site.

#2 @Lee: Although I think you're right, I don't think GG will update the vehicle class anymore. If I recall a thread about this right, they stated this already (don't remember who/where it was, I think somewhere in the TGEA forums).

So as a workaround I would suggest to use the player class as a base and create a new vehicle class out of it. I heard from GG members that went down this road and had success. So might be an option for you.

#3 Lee- it's a pity the collision side of things are so bad, because the actual driving physics are quite good, if treated carefully. I'll see if I can document the multitude of hacks and fixes I put together over a period of about 6 months last year.. I have collisions working more-or-less fine now: very few 'weird' results and no crashes or freeze-ups.

#4 Thanks for the responses.

Tim: I hear ya, but still...

Martin: Not a bad idea, I may well do that later if I get time/resources, thanks.

Sam: That would be very very cool, and I'm sure I'm not the only one who would be grateful!

Well, I'm a bit disappointed no one from GG deigned to respond. I guess the required 4 second GG splashscreen on my game will be a nice advertisement for a great locking up game engine! I mean come on.

Like I said...it's fundamental. They've got this great community they were wise enough to encourage, and I really suspect that it's only recently that Torque has become a really, really full featured engine when you include all the wonderful user-submitted resources. I don't think I could have done my project by myself any other way, and I see some other pretty darn cool looking games in the pipeline.

Point being, Torque's profile may soon be rising soon thanks to the groundwork GG layed several years ago. I mean, is there any other engine with so many options all ready to go?

So it'd be nice if it didn't lock up.

#5 @Lee:

Tim Heldna is correct, GG sees the vehicle starter kit as a bonus. If you are making a racing game from our very simple vehicle starter kit, you will need to understand vehicle physics enough to finish it up. We provide the starter example kit as a starting point, it was never intended to ship commercial products. GG could spend all of its time creating and selling a vehicle physics simulator/racing specific game engine. However, that is beyond the scope of a general game engine,a dn we do not intend to do it.

There are commercial uses of TGE with car racing games, the most recent of which was developed by Disney and is on display at Epcot center. I can also tell you that the commercial team that created the product fixed the racing physics for their particular needs.

#6 @Jeff: I appreciate your replying, but I've got to say that is as big a load of bull as I've ever heard for an excuse for not fixing a product. And I should know, having handed out my own fair share for the same purpose. :-)

A) It's not a vehicle physics problem, it's ALL rigid body physics, a core component of your engine.

B) If it was never intended to ship commerical products, why did you sell me a commercial license for 1.4? I'm pretty sure it said "commercial" and was significantly more expensive than the indie license. Mind you, it was quite fair, even with this defect.

C) Please don't tell me it's not a part of your business plan that at least some small percentage of your more casual users should eventually develop into more significant business partners. That has always been my intention, personally. Not trying to yank your chain here...haven't got any money now...but I will.

D) The last inference one can draw from your statement is that "commercial" users are expected to patch this problem for themselves, correct? Thus, what you give us is the same thing you give them, and thus intended for commercial products.

Look, like I said...I can forgive all kinds of bugs, and I certainly do in the case of Torque. But this is a category 1 defect, and with all the really talented people I know you have there, I just feel like it should have been fixed by now, considering how long it's been a known defect--at least not to get caught in a runaway condition or crash, if not ultra-realistic physics. I mean six years of development have happened, and you've added a whole lot of value to this product, but neglected this monster that affects such a large percentage of customers?

Bah!

#7 Well there was a total fix but created by Akio integrating ODE physics into the engine to create outstanding vehicle performance. I know, I tried it.

www.garagegames.com/blogs/27965/9393

He was told not to sell it because it violated the EULA, which in the method he was trying to sell it, was true. Unfortunately it seems he hasn't been around since last December and he removed his email link from his profile. I say this because he did offer to make the changes as hired programmer.

Here there is more discussion on the integration, but there are somethings I would like to ask about that are related to each of these threads.
www.garagegames.com/mg/forums/result.thread.php?qt=28247

In both areas it has been mentioned that Garage Games has been working on physics that might involve ODE integration or some other physics engine. It has also been mentioned that there was a possibility of a physics pack to be released.

Are these just rumors or is there truth to them?

#8

Quote:He was told not to sell it because it violated the EULA, which in the method he was trying to sell it, was true.

Personally I think this is the biggest limitation the Torque engines have - the EULA paragraph 3. It stops *any* programmer from making code packs or improvements and sell this stuff on his own. I think a lot of code packs never appeared or will not appear due to this limitation.

I heard the reasons for those limitation and it's GGs decision if they want to go down that road, but personally I don't wonder why noone has fixed that vehicle stuff or has released a new vehicle class. Reason: If you don't get GGs permission to sell (which you usually would get only if they publish it) you never can sell it. Period. So why should any programmer invest the time to make something new and cool if he can't make a few bucks out of it?

That is in my eyes the reason that so few code packs appear here. I'm pretty sure the demand is *much* higher for code packs than what is actually served.

So, regarding Akido, I just wonder where those talented programmers like him have left to? Does anybody know which community they joined then?

Slightly OT, but needed to say this. Not meant as flame.

#9 I am curious how the Torque Lighting Kit got to do it? It was definitely a code pack.

As I mentioned before Akio has pretty much dropped off the map. Before he did though he had mentioned he was willing to be hired as a programmer for projects to modify your code to use the physics.

Just a thought, but perhaps GG should have hired him to do the work, then they could have distributed it as an addon pack. They would have made a killing.

Nevertheless, what is done is done. I am not trying to flame, just recounting the past and what could have been an incredible enhancement for an engine that only uses rigid physics when so many others have ragdoll and much better vehicle physics.

Could have started charging more for the engine if it was included, I sure would have paid.

But all of this brings me back to my question about a rumored physics pack, is there any truth to it?

#10 You can sell code packs... there have been plenty of them. The limitation is that it must be published via GG to ensure that Torque source code isn't released to non-Torque owners.

#11 For what it's worth:

Breavetree's tank pack derives from player, not vehicle. This is a hint for anyone who wants to build their own crazyvehicle.

There's an ODE Physics pack *shameless self promotion* here. It's still work in progress, but it works for many many things. It's also free; if it doesn't work, you get a full refund without even having to contact the vendor.

Torque's collision code has problems, it's true, but my physics code has so far been working out fine with builtin TGE collisions. www.youtube.com/watch?v=gjEeN5er6aM

Gary (-;

#12 Additionally, this is in 1.5.1. Might be in 1.5, not sure. But it supposedly fixes most of the popular complaints.

Gary (-;

#13 Yes yes, like I said, there are code packs that come out, but by far (in my eyes) not enough to serve the demands here. And yes, like I said, code packs only published through GG. But exactly that is the drawback in my eyes. If GG for some reason does not like the pack it will not show up - regardless how good it may be for many others that actually would find this pack useful - no chance to get it on the market. That's what I dislike. Not all ppl out there need super polished products / code packs. Sometimes less is more.

#14 Hey, Martin Schultz is your pack still coming out. I check the site almost everyday to see if it is up.

#15 @Fucifer: Yes, it will come out, definately. Promised. And not too long in the future. There have been some... "difficulties"... around and I'm trying to master them all now. I'll drop you an e-mail right now.

#16 You know, if a new physics pack was being worked on by GG or an affiliate, I would personally consider that a legitimate reason not to bother with the old stuff.

It would make me feel a lot better to know that, though, for what it's worth.

#17 OK the question has been posted, someone wanna shake one of the Garage Games guys for an answer?

Don't get me wrong, I think their stuff rocks, but this is a pretty old problem and the only one I have seen really come out with any sort of collision enhancement stuff in a while is Ben Garney.

#18 Okay, I've got 2 things to add to this discussion. The first one is positive.

1. You can fix the Rigid class. If I did it, so can you. All the work I did on Rigid was done before I ever even opened a college-level physics text. My current implementation not only never crashes or freaks out, it allows you to perform expected (but disabled in stock TGE) actions like pushing one vehicle with another or knocking boxes over by rolling logs down a hill (open rigid.cc and you'll find that there's a function for this; making it work takes a bit more effort).

It's not 100%, and by that I mean that it completely lacks dynamic friction (in other words, if you move the box that a second box rests on top of, it can slide out from under it because it only transfers collision force [in this case upward motion], not frictional force [side to side in this case]). That said, it's completely functional for multiplayer vehicle interaction physics, or even the occasional physics prop.

If your next question is, "when will this be a resource," the answer is I'm not sure if/when. I'm not entirely happy with the methods I've used to get this working, regardless of the fact that it works extremely well.

2. I would guess that it's entirely possible that someone at GG has played around with PhysX integration, or maybe ODE, but the last thing you want to do as a professional developer is start copying things from your "good ideas" list into your "planned features" list. The instant one of them comes out and tells us that they've been messing with some physics engine in TGE, it gets turned into a grand announcement by the community; something along the lines of "GG Plans PhysX Integration in 1.6!" A few months later, when 1.6 rolls around, people are outraged that it doesn't have the feature, and the company has no idea why anyone expected it in the first place.

This is the same reason you don't give dev builds to the public. People freak out, and they create hype over things that may or may not happen. I think that, considering GG's position on this issue, you can assume that there is no official project to update the physics at this time. That doesn't mean that it's impossible that work is being done on it, but it does mean you shouldn't get your hopes up.

#19 Yeah but that's no reason to spout some nonsense about "it's not intended as a commercial product" + "you're too dumb to solve this yourself, that's your problem". A straight-up "we just don't have the business resources to commit to it" would have been preferred, a la Jeff's Linux announcement.

I mean, we're all grownups, here...well...okay, YOU GUYS are grownups and I promise to act like one for the time being, okay?

#20 Who cares about code packs. Hire the guy for $50 to add the code to your project. Done. Easy as pie. There's nothing stopping anyone from making money from their work with Torque. You just can't make easy money from advertising a piece of code work for anyone with/without a license to buy.