What would you like to see addressed in Torque 3D?

#1 1. Integrated PhysX, with Ragdolls.
2. Fully integrated AFX, not a separate purchase.
3. Voxel based terrain system with overhangs, etc.
4. Built in swimming, crawling, crouching, climbing, hanging from ledges, etc.
5. Built in inventory system.
6. Built in conversation system.
7. Built in Melee system.
8. Built in SSAO
9. Built in full screen shaders (blur out, etc)
10. Built in Radar
11. Square replicators
12. Replicator blockers
13. Built in terrain material references www.garagegames.com/index.php?sec=mg&mod=resource&page=view&qid=15075
14. Better default player artwork that doesn't use the wierdest starting pose ever
15. Ability to retarget dsq animations to different skeletons at load time (by providing a reference skeleton).
16. Fix the water bugs - Water doesnt' work right at mission load. You have to force it to resend the network packets after load.
17. Fix more water bugs - If you turn off reflections, the cube map is wrong.
18. Fix even more water bugs - You get incorrect reflections / blobs on the water seemingly from nowhere.

#2 @Matt

I am making my list now:)

But seriously and I have to go back an re-read the blogs on what Torque 3D is offering, because honestly I have not followed it that closely. I do believe there was mention of a Physics integration, perhaps Bullet. I am not Physics expert but from what I have been reading up on PhysX seems offer more features. I would like to see supported integration of PhysX into T3D.

When I say supported, I am not saying it has to be included, but at least a GG supported documentation (Mitch) of how it can be integrated.

#3 I think Jaime hit it good. Alot of the high points.

A fix for all basic Vehicle classes (wheeled and flying) and related physics.
You named one of the items with the flying vehicles crashing when moving to fast. Thats something thats been a bane ithink since the engines been around. With the game crashing with vehicles moving over a certain speed.

How about some more Ai choices and options. ie Aiflying Vehicles and such.
AI player is nice, but how about adding in Ai Gaurd for the new folks as working examples, yes it can be added to but it seems that aiplayer and aigaurd/patrol would cover a few of the basic requests.

oh, and i had a few more

Can we get a working movie player for in game movies.. we had Theora but its not compatible. And can we fix the demo recording and playback functions?

#4 It's banal, it's boring, and it's irritating to keep up, but:

I'd request an example implementation with one of each type of asset the engine is supposed to support both for internal and external use. Throw in an atlas, a player, a hovervehicle, a flyingvehicle, a rigidshape, ect ect (or their equivalents) and make _sure_ they all work. Or remove what doesn't off to an extention-pack to be worked on when resources are available to look those aspects over.

It slows everyone down to back-step supposedly functional code and figure out at what point the inheritance broke down when there's nothing there you're testing for, and have to cook up a test case from scratch. New tech is nice. New tech that doesn't break old tech however, is far more important.

#5 Linux Support.

I don't really hold hope of this, but I just want it noted that Linux support is the difference between me buying something and me not buying something.

Gary (-;

PS Then Plan9 support, but it's ok to put that down nearer the "Blue sky" section :-)

#6 I have to agree with Jaimi's suggestions and Edward's AI suggestion.

#7 Integrating AFX and PhysX and stuff shouldn't happen. Some of us don't use them and don't want them. It still shits me that someone thought it would be good to integrate the lighting kit with the latest TGE.

I'd say, scrap a lot of the existing stuff, and add/implement:

- Much more modular system for add-on game components and capabilities.
- Removal of old poorly object-oriented Tribes 2 game code in favour of a component-driven framework.
- Implementation of a single player version of that game that doesn't require having multiple server/client objects and partial networking to run.
- Decent lighting system that doesn't take (what seems like) hours to light a mission.
- Working linux dedicated server.
- Proper usage of namespaces in the code. Put everything Torque-related under Torque:: for example.
- Changeover from 3 spaces per tab to 4 spaces per tab (ha!).

Probably a lot more but I've lost interest in favour of Left4Dead... =)

#8 Some small things Ive always wished for...

1)x,y,z spinners , so I can edit a DIF/DTS position without having to type it in all the time.
Similar to 3dsmax.

2)A material system that can update itself and have a gui interface in real time without having to reload
and edit a text file save and reboot everytime, similar to the Plastic tweaker.

3)Elevators, moving lifts that affect players position similar to
Ramen Sama's Platforms that Players can Ride on.

4)If you create a Shader in Render Monkey , you could use it as is without
needing to know how to edit the Pixel or Vertex Shader File.

5) World editor/Inspector/Terrain editor..etc windows become floating windows so that they can
exist outside of the Game window.

#9 @Daniel - Nothing says you have to use them.

#10 Stable basic ground and air vehicles!

Total bane of my short Torque life and the reason it's currently side-lined. I know it's fixable as Luma and others have demonstrated, but trawling through the forums only leads to various failed attempts to improve things for the wider community. Basically the collision issues need looking at.

Specific issues that I can recall:

Vehicle becomes embedded in terrain/interiors when colliding at speed.
Engine "stalls" for a second or two when colliding with a RigidShape at speed.
Occasional crashes on collisions.

I can try gather together all the links and references I had relating to vehicle issues if it'd be any use? I think most of it is covered in a forum thread or two dating back years though.

#11 I'm going to have to second Jaimie and Edwards posts. Those are awesome suggestions IMHO. Also I want to throw in a request for the ability for Torque to have its own "Export to Web" ability much like unity 3d or dxstudio's web players. And to be able to get the effect of mapping a seperate detail texture to each terrain texture layer so that dirt can look like dirt while grass can look like grass etc.

#12

Quote:
15. Ability to retarget dsq animations to different skeletons at load time (by providing a reference skeleton).

I'm not sure I follow you on this one...do you mean something like in the current T3D demo where it shares one set of DSQs with all of the different Players (except for the Elf who has animations built-in)?

#13 Would you guys be interested in hearing my thoughts on your suggestions? Even if I say that something isn't likely to happen or if I disagree with you on something?

I don't want to get into a series of running arguments over a bunch of issues but I am more than happy to sit down and explain my thoughts on these various suggestions like I did over in this thread.

Also, I am willing to start talking about some of the stuff we have planned for Torque 3D in more detail but you guys will have to understand that they are planned and not already done and are thus subject to change or getting cut. We really don't want to get into another "TSE unfulfilled promises" situation where things are taken out of context and used to beat us about the head for years to come =P

I can promise that if you are mature and respect me then I will treat you maturely and respect you.

Finally, I am absolutely 100% willing to listen. I have no illusions that I am the smartest product manager on the planet but at the end of the day it is on my shoulders and on my teams' shoulders to get this stuff actually implemented and what we say, goes. If you find yourself taking it personally or are unable to deal with agreeing to disagree, then it might not be a good idea for you to participate. I'm not trying to be harsh...just trying to set some clear expectations ahead of time to try to avoid pain all around.

#14

Quote:Would you guys be interested in hearing my thoughts on your suggestions? Even if I say that something isn't likely to happen or if I disagree with you on something?

Wouldn't be much of a forum with one way dialogues going on all the time. Nor in the slightest constructive. By all means.

#15 I'd like to see an editor as in tgb 1.7.1 with a behavior system (of course in a 3d environment).
You just "drop in" your 3d Object and click to assign your custom made behavior to any of the objects.... now that would increase production speed drastically.
- Michael

#16 1. Pain-free, 100% correct, basic, static model imports. Any modern game engine/platform should be able to load a .obj, a .x, and a .3ds model into the engine for rapid prototyping with one line of script or a simple dialog box.

2. A modern 3D model format that doesn't have fundamental limits because of 16-bit indices.

3. Edit/Compile/Run development supported on Mac, Win32, and Linux. No more excuses and whining about Linux -- pick a popular distro, and make it work.

4. The ability to disable/replace *all* client/server/networking/ghosting abstractions for single player games, or for games with far larger numbers of objects in motion than the TGE/TGEA architecture currently allows.

5. A means to use someone else's physics implementation (eg, ODE).

6. A true scenegraph, not a bag of objects. Any node in the graph should be able to be a parent to any number of children, and those children can have children, and so on. Panda3D has this nailed.

7. Python for scripting. Torquescript should become a backwards-compatible mechanism for legacy code.


Basically, one's first experience with TGE/TGEA rapidly turns into a depressing battle with an exporter or quirky editor to get *anything* that's not a skybox or a simple textured cube into the engine... followed by a great deal of head scratching trying to figure out what script is where and when it's run and why someone thought it was right to put it there...

#17 Basically I need all the features that been done by the guy in his spare time as he stated in the NextGen TGEA Gaming thread:www.garagegames.com/mg/forums/result.thread.php?qt=79304. Why? Simply because it's been proved doable. The most important features of his work, in my point of view, is the FX material system and the HLSLFX enhancement which lay a solid foundation to quickly import thousand of shaders already made.

After all, the most noticed feature of a modern game engine is the shader driven architecture. Both graphics and physics can be implemented and should be done by that in the near future. Otherwise, Torque 3D should provide a WYSIWYG shader editor framework like other high end game engine been done before and I don't believe that will happen soon. Last but not the least, the muticore support using OPENMP also proved doable in his work. It's grateful to hear that those features will be available soon and I'm waiting to see it officially announced!

#18 Rather than a ton of build in features, I'd rather see a good solid core product with a modular way to integrate additional features to the engine.

Some of those additional features may ship with the engine as free modules, whilst others may be products in their own right. Kind of like the way you can include/exclude certain modules via the TGEA project generator.

Quote:7. Python for scripting. Torquescript should become a backwards-compatible mechanism for legacy code.

Again I'd rather see TorqueScript as the base scripting feature, with a mechanism to replace it with an alternative scripting language. Rather than forcing a specific language. The code needed to get those alternative scripting languages to work with the abstraction layer might be free resources, community resources or full products in their own right. Likely to be easier said than done, but I think it would be of greater use than just changing from TorqueScript to any other scripting language.

#19 Just get an engine with a better terrain system and foliage replicators because, i hate the current foliage replicator. i looks like wolfenstien 3d!

and maybe some better camera features, like have the viewport run from and object not a node

#20 @Jaimi
I do kinda feel 4-7 and 10 are more "mods" than "base engine" -- not that I'd turn them down, I have to say.

@Matt
Built in batching for objects, like ShapeReplicator but being able to choose exact positions rather than use random seeding. (though there might be a very good reason why this can't happen)

Materials on terrain.

Constructor export/ in-game performance issues sorting out.

Jaimi's list.

Er.... "the moon on a stick" ... a "build my game button".