Canardia

Then you need to make sure that Leadwerks.NET is fully documented, if you want it to be a candidate for official support too. It might be good to have in mind that LEO could be developed to be equal with Leadwerks.NET also, or then LEO could be also dismissed and Tiger would be the equivalent to Leadwerks.NET. Basically the difference between LEO and Tiger, as well as between LEO.NET and Leadwerks.NET, is that they are more clean OOP structured, and not using so much procedural types and structures. For example in Tiger there would be a real Vec3 class, which is probably what Leadwerks.NET also has. This would also mean that both Tiger and Leadwerks.NET are actually closer to the BlitzMax mod, as it has also a real Vec3 class. This would make it actually easier for Josh to support it, if it matches the BlitzMax syntax, and is very close how the engine internally really works. And it would be closer to LE3's syntax too, so the C# headers could be easier ported to LE3, and then LE2 and LE3 would be more similar also. Only benefits would come out in my opinion.

I would suggest to Josh in order to qualify the C# headers as official the following should be done: 1) Verify that LE.NET works exactly like the procedural C headers. 2) Verify that LEO.NET works exactly like LEO, the OOP C++ headers. 3) Keep Leadwerks.NET outside of the official support, since it's like GameLib or Tiger which are also community supported. 4) Name someone who wants to coordinate all these steps, and who serves as contact person when there are problems with the LE.NET and LEO.NET headers. Then Josh can contact this names person when someone reports a bug with the C# headers.

I don't think Josh can support much with the C# headers as the doesn't know the language very well. You just need to make it working and test it, then Josh can at least include it with the SDK. I think the problem with the C# headers was always that they had too many bugs, and nobody actually tested them fully. That's really the only difference to the C/C++ headers, which are always tested and very carefully updated. The other only difference is that the C# community doesn't have a clear infrastructure. You need to elect someone who is responsible for the official C# headers. I think it was ZioRed earlier, not sure if he still is, or if he forgot about it It doesn't mean this guy needs to work 24/7 on the headers, but he just needs to organize the suggestions and votes and keep the official statements updated. Josh can then support the C# headers using the official statements, and if something doesn't work like officially stated, then ZioRed gets spanked. If ZioRed, or whoever the official C# coordinator is, doesn't want to be the official coordinator anymore, then someone else gets elected. That's how it goes with the C/C++ headers too. From time to time Roland is the official C++ guy, and from time to time he asks me if I want to be the official C++ guy again, since he is too busy with arts or other things in life. But there can only be one, at any time. Like in Highlander, Matrix, and StarWars too

Why don't you guys just forget about the words official and community and just do it? Call it tested and fully functional or something like that instead.

Why don't you just make a seperate get/set and and a GetValue()/SetValue() header version? I don't know how good a C# compiler is in optimizing and eliminating unused code, but in C++ you could have both in one file, either using a #ifdef or just letting the compiler remove unused stuff. There are people who prefer get/set, and there are people who prefer GetValue()/SetValue(). And they will keep their principles. Both methods have their own benefits.

CreateFramework, RenderFramework, that was introduced in LE 2.32.

Yes, and for networking you would really want the maximum speed which only C++ can deliver, so a BlitzMax networking solution would be rather useless anyway. There are certain things where you absolutely need maximum speed at all costs, and networking is one of those. AI is another. BlitzMax is fine as long you use it only to send stuff to the GPU, like LE 2 does.

Yeah, but when people realize that nothing is a fact, they start searching for something more interesting. Then they find different people's opinions as the next best thing after those false facts. After that they might even find the people itself interesting, or even better: the things they make and say To understand this, you have to watch this video. Life is just a ride: http://www.youtube.com/watch?v=Q95kX_EP2Nk

That is actually a fact that everything every human says is not a fact, but only their subjective opinion. So saying "this is my opinion" is kinda useless, since it would need to be added to every single sentence anyone ever says. We don't even know for sure that we or anything else exist, we might just be simulated by some god. Some 13 year old god playing simcity with his universal computer

LOL, it's just a hobby. Actually I do much more for my real work!

Yeah, Max3D was long time the #1 engine, but it was never finished. But it had such good features that it even beat LE2 at that time. So my imaginary engine will indeed have some features from Max3D

Actually I'm not sure how realtime the Editor in LE3 is. I thought also at first that it was not realtime, but then I read that it's not realtime for some part only which nobody needs anyway.

But it works. People talk that it's wrong, but nobody can actually point out WHAT is wrong. I say it's not wrong If someone finds an error in the chart, I will fix it immediately.

Nope, I think all the criterias are of great importance, and I've explained there also why. A criteria importance weight system is also a bit difficult to make, since different people have different importances for the criterias, so keeping them all equally important is not the worst choice. A people's opinion importance weight system would be by far easier to make For me, as musician, OGG is VERY important. I can literally hear the difference between a MP3 and OGG song, and most of my songs don't even work as MP3 because they have very high pitched sounds. MP3 just removes them, no matter how high you set the bitrate. Sure, WAV sounds as good as OGG, but also not always, since a WAV file can be HUGE and thus cause stuttering in the playback because it doesn't fit into the memory and the CPU can't keep up with the streaming speed of the huge unpacked data. And besides, OGG packs better than MP3 and lossless, and disk space is also important for musicians and gamers and everyone else.

I could put some imaginary engine on 1st place, just to calm down the prejudging people a bit Actually that would be a nice idea, to put on 1st place an engine how I think a engine should be.

I have already made a engine comparison sheet: http://siipi.com/public/mika.nsf/blogs/2AAC3EECC888AC57C22575F7004ADC16 It needs to be updated a bit though, but nothing spectacular has changed since. Btw, Leadwerks Engine was long time on place 3, but after that it has been on place 1 for a few years. So it's not a biased LE fanboy chart as many would assume at first sight. And I actually own 20 of those engines too, but I only use LE because it's the best

Have a loading screen which checks the FPS and when FPS is over 30, then close the loading screen. A more advanced way, which needs no loading screen, would be to set first the mass of all physics object to 0, and then set the mass to 1 in regular time intervals, like every 10 second activate one gate. Since the player is not near the other gates, he wouldn't notice it either, so make sure you start to activate the gates from nearest to the player.

You can just export from UU3D with only bones and animations checked in the export dialog.

You don't need to put the animations in the model, but you can have seperate animation models which contain only the animation. This is useful when multiple models share the same animations also. Then you can use the LoadAnimation() command to load the external animations. You can also assign physics bodies to the bones, and let the physics engine do some part of the animation. In addition you can create physics joints and move the model like a marionette or robot.

LE 2 has big enough terrains, and it makes no sense to use big terrains anyway, since then you have tons of stuff in memory which you don't even use, and the loading time is huge. So LE 2 doesn't need bigger terrains, but positionable terrains. You can build a infinite streaming terrain scene out of small terrain blocks which are positioned when loaded and then removed from memory when they are out of range. I'm testing next how it would work out when using meshes as terrains. It should be possible to use Blender's sculpt tool to create the terrain meshes, and then you can have also real 3D terrain with holes, loops and overhangs, and there is no limit on textures.

I'll just put it here, I think it's good for people also who are considering buying Leadwerks Engine. Here it is: http://www.leadwerks.com/werkspace/page/Documentation/LE2/tutorials/_/programming/quickstart-tutorial-r69

While the C headers are officially supported, LEO is kinda in a hybrid situation. It is officially supported, as in being part of the SDK, but for problems with it (there has been none since it's perfect) you'd have to ask in the Forums. Roland wrote LEO originally and I continue updating them when new engine features come out. The lack for official LEO support was partly also the lack of knowledge of C++ of the officials, but with the LE3 development the officials have gained huge C++ knowledge, indeed even to a state that they love it So LEO should be theoretically now 80% officially supported.

Well, but .lib is C++ only also, so there could be as well a OOP DLL for C++ only, since compiling a .lib can take very long, and compiling with a DLL is instant as it's used only in runtime. You have to remember the majorities of developers also, I think nearly 5 billion people use C++, and the other languages have only few percent share total. Here are some statistics, which of course don't tell the whole truth as each site measures them differently, but clearly C/C++ are always on the top. No idea what that Java is doing there, probably because the insane amount of LC indian developers: http://langpop.com/ I personally prefer the Freshmeat and Google Code statistics, as they are reliable and meaningful sources. Lambda is also interesting, although I find Haskell quite horrible, but it's a new experiment whatsoever so it gets some attention before it wipes out. Haskell is very similar to Fortran though, and Fortran is faster than C++. Sad but true

Actually LE is Object Oriented, but only the DLL is procedural, because BlitzMax can not export OOP DLLs. So you can go and reproduce the original internal OOP structure in Leadwerks.NET like it was done in LEO also. LE3 will have much more library forms, including native C++ .lib format and possibly also the OOP DLL. The procedural DLL form is still needed for languages which don't support OOP.

I can't upload anyting in the Forum, it says: "This upload failed".