V12 Features and documentation

#1 Joe,

Maybe I missed it, but where's it say we get the source code? Far as I can tell, it's just a library that you can link with. That doesn't leave much for developers.

5000 symbols? My 3-week-old demo has more than 5000 symbols! This is not much, really.

Their feature list looks pretty good though, and the screen shots are nice. But a comparison isn't possible with the level of detail listed on that page. Seems like a good package overall (and you can't beat free!).

I'll definitely keep an eye on it.
--Bryan

#2

Quote:
If you were to use Jet for a commercial product, the current proposed fee is US$ 30 000, payable at any time prior to the release of the product.

This is one of the biggest differences. The fee for the V12 for a commercial or non-commercial product is $100.

Quote:
Will AURAN Jet support multiplayer networking?

Yes, however, the developer will be responsible for the implementation of this feature.

V12 will support multiplayer networking right out of the box.

Quote:
Does AURAN Jet support the ability to have either random landscapes generated by functions (fractals) or by generated tiling?

Landscape tiling will be supported when the map editor is released with a future version of Jet.

This is one of the strongest features of the V12. The V12 will ship with an excellent terrain/mission editor.

With the V12 you get the full source code for the engine but with Auran Jet you get a bunch of libraries and the source code for examples using those libraries. So if you wanted to add another renderer (Glide for example) then you wouldn't be able to with Auran Jet.

The V12 will support building structures in Worldcraft, terrains in its terrain editor, and models in 3D Studio MAX 3.x. Auran Jet only supports exporting structures, terrains, and models from 3D Studio MAX 2.5.

If you use Auran jet you will still have to find a publisher for your game (very hard to do) but with the V12 you are guaranteed a publisher =)

The documentation for Auran Jet seems to be far more extensive than the initial V12 documentation will be but the V12's documentation will be expanded rapidly after it is released.

I have only done an hour and half of research into Auran Jet but these are some of the obvious differences that stands out.

#3 Kewl. A hundred bucks and royalties for just code is pretty not a great deal. Maybe garagegames should put more work into their documentation. Maybe they should use automatic code documentation tools like Surveyor http://www.genitor.com/genitor/whatsnew_Sur45.htm or http://www.doc-o-matic.com/

#4 You get the code and a demo program (with code). There will be a little documentation in the initial drop and a lot more will be added quickly.

#5 The main problem of Auran Jet is performance+supported features. I haven't tried the new beta, but the beta_1 was a total putoff.

#6 Yer rite. Tha beta 2 wan not much beta, and it really sucks ass you have to rite your own tools from scratch for 30,000 dollars.

#7 With my TNT2 I recieved 30+ FPS with 15,000 polygons in Auran. Without progressive meshes either. It was a small bit jumpy though, every few renders..
It was a one object test.

Auran seems powerful, I hope V12 is as powerful as Auran.

>Auran Jet only supports exporting structures, terrains, and models from 3D Studio MAX 2.5.

Not true. It has had a 3.0/3.1 exporter for a while.

People were debating the fact a lot in the Auran forums that you can't have a better editor than 3ds max made by a graphics engine, so having to re-align your objects in their editor and have to export one by one to acess objects differently would be a waste of time, effort, and resources...
You can even program engine/game specific tasks/settings inside 3ds MAX if you write your exporter well.

Anyway, Auran is good, I hope V12 is better.

#8 hmm. when I tried the beta1, it defintly not that fast. seems they optimized a bit since then. anyway, depending on the scene and you CPU power, but 450,000 tri/ sec isn't all that high.

#9 I have been playing around with Auran Jet since beta 1. Performance definitely improved with Beta 2 - just from experience I never gathered any data but it was pretty obvious.

I certainly have a few gripes with the engine but I don't think its really fair to compare it with V12. You have to keep reminding yourself that Jet is still in Beta and it is not intended to do the same job as V12. The lack of documentation (and there is already enough to get started, even for a rookie) is also not surprising for a Beta.

Also jet is very open ended so not getting the source is not that big a deal (it is a shame though). On the other hand the guys at Auran stick to the "must not be game specific" policy far too rigidly. They use it to justify the lack of tools "a level editor would be too game specific" and the lack of features "a terrain system would be too game specific". One other thing, Auran don't communicate as well as GG, but they are still better than most. [disclaimer: these are NOT direct quotes]

If you are making a FPS I can't see any reason not to use V12 over Auran. But for other 3D environments I think its worth waiting until July to see how RC1 goes. But remember RC1 is still not the final release.

I also think that GG should be talking with companies like Auran about possible cooperation - although I'm sure they don't need me to tell them that :)

#10 15,000 polygons is pretty damn good... Compared to many at least...

>The lack of documentation

Huh? Auran jet does have a lack, but doesn't V12 have a much worse one, especially for a nearly completed engine?

I think overall I will like V12 better than Auran jet...

#11 what kinda scene is that when you run the testing?? 15000 tri all visible? anyhow, I can easily push visible 600,000Tri/sec (20,000 tri/frame, 30fps) in game on my aging Celeron350/TNT1. so...ha.

anyway, are you getting your new computer tomorrow?? with a Geforce? if so, try that test on the Geforce see how much faster comparing to your P3-600/TNT2.. i didn't bother to DL the jet_beta2 as it ask me to register.

personally, I am not interested in getting these engines, I am more interested in using their performance as a reference.

#12 I dont know why you would even consider a $30,000 product over $100 product. And what's all the hoopla about documentation - I am not even sure I know what docs are :) As a developer, I am WAY more interested in the code than docs because the code ultimately is the final documentation.

But you never know, maybe a few docs you will use a couple of times are worth $29,900 - lol.

#13 Having extensive documentation makes using technology radically easier. Source code, even with comments, is a horrible form of documentation. At best, it can give you an overview of a single code module. At worst, it can lead you astray with incorrect comments or just be incredibly obfuscated.

Source code does NOT tell you how the pieces of a system fit together and interact. It does NOT tell you how the data flows through the system. It does NOT tell you any higher-level limitations you should be aware of.

One can figure all of these things out independently, but the amount of effort required is huge compared to having documentation readily available.

I expect that GarageGames is aware of this and will have quality documentation available for V12 eventually. In the meantime, I'd expect a lot of inane questions to flood the message boards as people come to grips with the very complex system that makes up a sophisticated game engine like V12.

And Justin, your argument is like comparing apples and oranges. Even if the V12 and Jet engine were identical in every aspect, The Jet engine's contract is far more flexible than V12's. You can shop around your Jet engine powered game to anyone. This starkly contrasts with the contractual requirements of the V12 engine. And this is assuming that the engines are equivalent. For all we know, the Jet engine could have Doom3 quality graphics at 60fps on a Pentium 100. This sounds ridiculous, but my point is clear - the final version hasn't been released - and you should wait to judge it until then.

#14 >i didn't bother to DL the jet_beta2 as it ask me to register.

I didn't register because it wouldn't send me the confirmation e-mail, so I got the URL from a friend and downloaded it anyway.

450,000 tri/sec was good for me after previously working with a slower engine...

I don't see how an engine could physically be that slow...

#15 I agree, nothing is better than good documentation. A step by step manual is the best, and could stop tons of new users from asking stupid questions that everyone knows who is a little bit experienced with the engine.

If you want tons of people asking stupid questions, then sure, have no or poor documentation, but if you want people to learn fast and learn the tricks that they need to and the ways to do what they want to right, then you need good documentation.

I like documentation that will take you into creating a small application, explaining how and why to do everything, and then in the end telling you how to expand the application, and then the other documentation is easy enough to use to simply add to it.

It makes it so much more simple to learn an engine.

#16 I will agree that documentation can sometimes help provide an overview of a system. But I have to point out that code IS documentation of the game logic. Most developers spend many hours re-documenting their code in a prettier format - it's amazing. You have to learn the code in order to manipulate it anyway, so why not get dirty and learn the system once. That's just the way I would do it, so I am not worried so much about V12 docs at the start. I understand that others may feel differently about the issue.

As far as the engine debate, most of us don't have 30,000 to spend for the ability to shop around their game. In fact, I might say that having GG deal with the mess that is publishing your game for you is a great benefit of the deal - not a deterant.

Having the full source for V12 and having GG worry about publishing your games - what else could you want? Oh yeah - I guess the fact that it cost $100 for GG to do all this for you is another amazing benefit.

As far as how well the engine runs or looks today, that doesn't mean much. Machine specs improve daily and the game will only run better with time - especially by the time any of us finish making a game with any engine. That said, Tribes 2 is being recognized today by the industry for it's amazing graphics so I wouldn't worry this issue too much.

In the end, it's all about your ability to design a good game - not how you present it. The V12 engine has an immense library of capabilities and is considered a AAA engine - all for $100. You simply can't beat it...

#17 Justin, you said, "Most developers spend many hours re-documenting their code in a prettier format - it's amazing."

I have to ask the question: Have you actually ever worked at a software company?

An abominably small number of developers actually bother to comment their code well and (even more importantly) keep the comments in sync with the changes to the code.

This is as true in the game industry as it is elsewhere - perhaps even more so, since software engineering principles have not really been embraced by the game industry.

One would hope that the developers of the V12 engine do clean up and redocument their code before they release it for public consumption, and I expect they will to some extent. But this is certainly not the usual state of things, especially for software that remains internal to a team.

While having the source code is great, it is no substitute for documentation. Ideally one would like both.

You also said, "You have to learn the code in order to manipulate it anyway, so why not get dirty and learn the system once?"

I'm not arguing here. Of course you'll probably have to learn the code eventually (see my most recent .plan update, which encourages people to do so). However, the task can be greatly simplified by a decent pair of docs; the source code for a large system is enormous, and you might spend days trying to figure out what a single module does only to discover it's not even used in the game.

Reading documentation does not mean you'll be learning the system twice. Good documentation is really only a roadmap; it gives you a high level overview of the system and lets you focus your attention on the things that are actually relevant. Would you expect someone to learn MFC by reading the source code? How about STL? ;)

The very thought makes me shudder. And V12 is larger and perhaps more convoluted than either of those systems.

Code is no substitute for good documentation. People expect to have tables of contents and indices in technical references; why don't they demand the equivalent from source code?

#18 As for the 'benefit' of having GarageGames publish your game, their contract is not a simple right of first refusal. They, and they alone, get to decide where or not to publish your game. If they choose not to, they will put it in the free games bin. You won't be able to shop it around to other publishers.

Using Jet means that you're not necessarily putting all your eggs in one basket - and you don't have to pay the 30 grand until you've actually published the game.

By the way, I'd like to point out that I do think the Tribes engine is probably still a better deal for most people. I'm playing Devil's Advocate here because I'm not a big fan of religious company loyalty and blanket statements like "I dont know why you would even consider a $30,000 product over $100 product." That's a gross simplification, and one that does not qualify as intelligent discourse.

Sorry to pick on you here Justin; other than that one statement, your posts are very reasonable. I just take issue with comments like that - it seems to be questioning the intelligence of people who are interested in the Jet engine.

#19 There's a fairly simple solution to the issue of GarageGames having ultimate publishing authority over your game.

Write a wrapper around their engine that insulates your game logic from the core engine functions. Then if you don't like their deal, all you have to do is replace the engine with a different engine, and remap the wrapper calls. Their exclusive licensing agreement only extends to games that USE the V12.

A game is more than its engine. Or at least it should be. ;)

#20 I doubt that writing a wrapper for abstraction to multiple engines would be easy at all. Every engine will have a completly different architecture, different features, the way that those features are implemented would likely be different, and different engines would have very different capabilities.

Even if you managed a wrapper for engines, it would end up being so abstract, and diluting the features so much, that you wouldn't get anywhere close to the full performance out of any of them. Also you would have to check if all the features are implemented or not. Engines do not all do the same thing, they are all written with a specific task in mind, ie the game that they were written for. V12 has quite unique terrain capabilities as far as I am aware.

I know there are probably exceptions (I can't say that I am familiar with that many game engines) but I doubt very much that a general wrapper could be made.