Torque's 2D and 3D Future

There have been a lot of hints and rumors about what GarageGames has planned for Torque in recent months and with this post I am ready to reveal much if not all, so prepare for a window into GG like you guys haven't seen in quite a while. This post will focus on our development process for Torque products at GG and our plan for some future releases in 2008 and beyond.

The Process

We've done a pretty poor job communicating this stuff in the past. We use a lot of internal references publicly that get misunderstood and Torque users start asking questions like: "When will Juggernaut get released?" or "Will Torque 2 support DX10?" etc. These are totally fair questions because, again, we've been less than clear. I'm approaching this post with the goal of making all such terms, especially "Juggernaut" and "Torque 2" plain and clear. I'll silently cross my fingers and hope for a temporary blessing of better writing skills than I generally have :)

OK...first up, let's clarify some terms. Juggernaut. What is Juggernaut? Joe Maruschak and a few others from GG have explained this in forum posts and comments, but I don't know if it's ever really been put in the big picture context. Juggernaut started out of necessity and has since become an instrumental part of our Torque development process.


Juggernaut began because we had 3 separate game engine codebases that all shared roughly the same fundamental pieces, but didn't share a core code repository. This was a big pain in the butt. Soon it became clear that merging these code bases was a necessary part of continuing to support the Torque line of products. Merging all these things, taking all the fixes from every game we'd ever worked on and every engine repository we had, and coming out the other side with a better, tighter, more stable engine would be no small task, as you can imagine. Nevertheless, developer magic prevailed and today Juggernaut is quite a machine.


Juggernaut's features at present still lag behind some of our current products. The performance is better but TGB 1.7.2 is overall a better product than the version of TGB in Juggernaut. There are more current features in TGB 1.7.2, but moving TGB 1.7.2 to a version on the Juggernaut code base with the same features would be a great improvement.

In the case of TGEA, version 1.7 actually takes the best of what's in Juggernaut and adds functionality to that (like multiple terrain tiles). The end-run goal here is keep all our products up to date on the same code base. We're moving very quickly toward that goal. TGEA 1.7 is the first step. It would be reasonable to expect similar updates to TGE and TGB to move those products (still on the older, incompatible code bases) to the new unified code base.

A unified code base has obvious benefits to both GG staff and to Torque users. For us, it's much easier to maintain and update Torque products. For you guys, there are more frequent and cross-product updates in addition to project portability. This means we can also do really cool stuff like share TGEA features with TGB. Want Torque X style shaders in your 2D game? Yes. We will be able to do this. Networking? Shouldn't be too much of an issue. There are a number of feature combinations that we can present in TGE, TGEA, or TGB, all of which add value across the board.

So this is all great right...Juggernaut is where all the the features of Torque live and play happily together. But there's more. Juggernaut is what our internal game teams use to create cutting edge games like Legions, and what external teams use to create games like Penny Arcade Adventures, Cyclomite, Metal Drift, etc. This means that code is constantly battle-tested in the most demanding conditions possible. It's basically used only by professionals *very* familiar with Torque who can operate with largely undocumented code, command line or very user-unfriendly tools, and the occasional quirk or instability.

We watch the use of features in Juggernaut very closely and gather feedback from the pros using it to improve performance and stability. Only when features have been through this process of shipping games do we gather together the fixes and features necessary for an update to TGE, TGEA, or TGB.

What about "Torque 2" ?

Speaking of poor communication...what does "Torque 2" mean? Is it a new engine? A collection of features being developed in Juggernaut to eventually be released in Torque? We have not done a good job talking about what this meant for Torque owners and I want to really help clear this up from both a process standpoint and a product standpoint. Here goes...

"Torque 2" does not mean Torque 2.0, TGEA 2.0, or any other version number of a Torque product. It's another case of an internal reference (like Juggernaut) being used publicly in a way that's been confusing. We talk about "Torque 2" internally to refer to our R&D repo. This is where new features are created and large refactors are done. When a change is too big or too low level to go straight into Juggernaut, it's handled in a separate R&D repo. One of things we do when brainstorming what sort of work should be done in R&D is say to ourselves, "OK, what do *we* need to make the kinds of games we want to make on the PC, OSX, Xbox, Wii, PS3 and InstantAction?" We take a broad look at the pieces of the engine we're least happy with and decide what can we do to improve it both for us and for Torque users. There are no sacred cows in this repo... things get shredded in the name of greater efficiency and feature sets.

A great example of a shortcoming that can only be addressed in the R&D repo is Torque's long-chain class hierarchy. It can be a big pain to work with and the consensus pretty much industry wide is that components are better. Moving from a fat class hierarchy to components is *not* a move that could happen in small steps and keep the engine stable and useful for us or for 3rd party developers who want to use our tech (i.e. Juggernaut). It's a long process, but big steps have been taken in R&D to swap our hierarchy for components. When this feature is stable enough to be useful for professional development, it will be put to use by our internal teams and 3rd party developers.

Another great example of a massive R&D change that happened in the "Torque 2" repo was Juggernaught itself... when we were first getting that started, it was a massive, fairly destructive R&D project and it's been built back up to stability over time. After the technology in "Torque 2" stabilizes, we move it to "Juggernaught", which is our pre-release repository and proving ground. From there, it gets used by us and other teams who aren't afraid of the lack of tools, documentation, and rough features in games, in order to smooth out the rough edges, fix bugs, tune performance, and all that good stuff. After putting a Juggernaught version though the rigors of shipping commercial titles, we go through the arduous process of creating tools, improving usability, and writing documentation for it. These are the final pieces necessary for a Torque product release.

So the process is this...


You'll note that there are lots of pieces from R&D that have gone to full production and there's much more in the works. This is a process that really works and ensures that you're working with the same technology that builds all Torque's best games. There's a lot to be excited about for future updates to Torque products. Speaking of which, let's pull the covers off our plans for 2008!

Opening up Development

From now on, with every opportunity we can, we're going to do beta releases on products so that we can get your feedback during development. This is just the first step toward a transparent development process that we talked about months ago. The TGEA 1.7 beta is the first time we've done this in a while, but not only do we want to do this every time we have a big release, we also want to find ways to get more of you involved. Right now, the open beta for TGEA is open, but before that, we also did releases closed to GG Associates. Associates are basically GG community superstars, but there are many of you out there who can help just as much and we want your input too. We're in the midst of building a web interface to support this kind of community development and feedback. If it goes well, you all will be able to use this system to host, track, and communicate about your own projects as well.


The impetus to open up development has been driven by another key component to Torque's success. You. There are tens of thousands of Torque licensees in the GarageGames community. A large percentage of you have been *incredibly* generous in helping us improve the technology. Whether it's guys like Jeff Faust, John Kabus, Melv May, John Kanalakis and Tom Spilman who create incredible products that extend Torque's capabilities or make it easier to use, or superstars like Josh Engebretson (of Minions of Mirth fame) who selflessly contribute improvements and bug fixes, all because they want the community to benefit from a better Torque. They aren't employees, but they're part of what keeps Torque such a strong technology and they deserve *massive* credit for helping GG deliver a game development better experience to you. So, to all those mentioned and all the rest of you that have contributed to our codebase over the years, thank you!

Torque Products

The TGEA 1.7 open beta here is going really well and we're looking at squashing 90% of the bugs that have been reported by full release. This is the best Torque engine that's been released to date and it's all FREE for current TGEA owners.

TGEA 1.7 is bound to ship in April and it's our best Torque engine to date. Just as TGEA has moved to Juggernaut codebase, so will TGE and TGB. That move will ensure greater stability and vastly improved performance. As new R&D is merged into Juggernaut daily, we'll be looking at new features to bring to TGE and TGB as well. There are some really exciting possibilities! On top of that, documentation and learning resources on all Torque products should improve dramatically. This means more updates to all the current Torque products before the year is over.

We've been listening and we know what you're looking for. There are exciting projects in the works here including porting Torque to new consoles, supporting new platforms, enabling new publishing options, and of course, lots more great games to inspire and motivate development. We're all going to keep working overtime to make Torque as great as the community that uses it. Again, *huge* thanks to all who've been exceptional, selfless contributors to the community over the years. You're the ones who make Torque great.

Finally, A Pre-emptive Q&A (with myself)
Q. Can I Buy Juggernaut?
A. You wouldn't want to. Working with Juggernaut as is can be very difficult and as of this moment, TGEA 1.7 is everything Juggernaut is. If you really want SVN access and to sign a crazy tightass EULA and NDA (because there's lots of stuff in there that can't be public...e.g. the Torque 360 code is in there and that contains proprietary MS info), then yes, maybe. It's a PITA for us and it would be expensive for you so nine times out of ten the answer will be no. If you become a GG Associate, you will have SVN access to our current Torque projects in development though and can participate in closed betas.

Q. Which of the releases for current products will GG charge for?
A. Not TGEA 1.7. Beyond that, we can't say. I would prefer if we established a more set time line for paid updates. TGE went years without GG charging for updates. TGB has been over 3 years. It really depends on what goes into the releases. Obviously development requires that we pay our guys, so at some point we'll have to charge, but as always, if we do, you'll get more than your money's worth.

Q. When will we see OpenGL support in TGEA?
A. Ahh...should have added that to the R&D bubble in the last graphic. There is work being done on it right now, but talking about a delivery date is still premature.

Q. How do I become an Associate?
A. There is no hard and fast rule of what qualifies a good candidate for Associate status. It is usually rewarded to long-time GarageGames members who show continued contributions of quality work to the community. Candidates should above all be helpful and willing to share their knowledge with the community through the forums or blogs. A professional attitude is expected when interacting with others in the community. Associates are looked up to by other community members, so it is important to present a good example of how a professional developers should interact with their peers. It is possible that some members who should be associates have been overlooked, if you know of such people, you can nominate them by e-mailing us.