Microsoft's XNA ?

Been playing with Microsoft's XNA Game Studio 3. It seems hard work compared to GameStudio A7 but easier than Direct-X. I just feel that this is where I should be going as you need to know alot more about "real" 3D code.

What do you think about XNA and its relevance to the real world of games design ?

Drac.

Honestly, there is little you can do in XNA that you can't do with 3dgs.

Define 'real' 3d programming. I mean 3dgs has a Z coordinate, that implies it's 3d to me

"real" 3d code that your referring too most of the time isnt even what they use to make the games, they use c++/c# to make the engines, and then most of the time a scripting language is attached to it, so in your head what you need to know is if you want to be an engine programmer, or a game programmer.

The XNA code gives people a bit perverted impression of 3d programming. In reality it is much more technical, not just mesh.draw(), etc.

Anyway, XNA is a nice package and simple enough for quite fast development of small and maybe even mid sized games. The approach XNA provides is quite practical, at least in my opinion.

However there still are some short comings, like no integrated dynamics (physics and stuff) library.

And even if you are a game programmer, it isn't really necessary to know the underlying implementation of 3d code. It might be useful but I believe that it would be a bit smarter to pick an engine and start making tons and tons of demos. Everybody loves to see demos, including the companies that are going to hire you in the future.

XNA is a kind of a hybrid case though. It isn't an game engine, but more like a framework for making an engine. A great tool for learning about making a game engine or making games in general.

As of relevance, I think XNA is becoming fairly relevant in the indie scene.

I agree with ambassador about the fact that Xna is more a framework than a game engine, than mean you have to create most of your tool or use 3rd Party. They should have physics integrated but I mean is not this relevant because Ogre which is a graphic library don't have any. + it's a free package. if you want to have for example a collision library you just have to pay 49$ dollars for the submission ( haven't did so I'm not really aware of it)

But if you want to compare both of them XNA is great for developpers you can build you engine in no time and also it give you a better control without work around. It's coded in c# so you can't find a lot of inofrmation ( a7 too but it's not the same thing) and also the community is growing, so you can progress.

A7 is a game engine so it's supposed to be a complete package, but it's not. I can really judge the actual engine since I don't own it( just have the old A6).

But the reason I stop using complete package is some point it's interesting if you want to make your game and everyting but are limited to the game feature, you just wait for the other to your job, you get lazy and you don't really progress. This is my reason and I do understand why people don't want to bother writting their own tool or classes, but since my game didn't really to what I had in mind Xna was a good deal beacuse I didn't have to create all from scratch but it same time I was free enought to take the direction I wanted.

why bother if it doesn't run on mac or linux.

there is little you can do more with 3dgs than you can with notepad either. it's not about i-could-implemet-it-features, it's about whats-already-in-there-features and how-fast-can-i-realize-what-i-dream-of. but this duscussion is pointless in this forum, i know.

with "real" game programming, i assume he means c++ and assembler, because that's what real guys like dracula do, program games in cpp and assembler

[quote=broozar]why bother if it doesn't run on mac or linux.

Maybe because it runs on Xbox. I mean cross platform is not relevant. Look unity is just mac but it's a great engine.
3ds Max as far I know is only pc and it's a great tool. So saying yeah this is a bad tool or engine because is not cross platform is just not the point at all

Knowing general 3d programming is pretty much a fundamental aspect of programming 3d games.

Demos are boring, if you really want to learn to code properly and not be a demo kiddie then write a game, any game to the end. You will learn more writing space invaders from start to finish than writing a demo of a sphere with per pixel lighting rotating in space.

actually, it is so not.

It is till unity 2.5 sometime in `09

the unity engine runs on mac, windows, iphone, and wii. who cares where the dev kit runs on? an xbox game surely isn't developed on an xbox either, but on a pc.

XNA is quite good if you want to use modern languages with OO Programming like c#...

Its fast enough for indie (and some commercial) games..
and the best: you can usw it within a great IDE - Visual studio.

A7 is also cool, but the SED is just out of date...
No code completition... A7 would be fantastic with a c# wrapper..
it would be the first thing i buy...

So if someone went to a games company and said, my experience is with the GamesStudio A7 game engine, and I also have a good working knowledge of C++, would they be taken seriously ?

Drac.

They won't be interested in which game engine you used. They will be interested in your C++ knowledge.

yeah .. but thats not the discussion here I think

brozar, have you ver tried XNA ? Because I wasn't really interested in it but since I knew c# from A6 I try it. The learning curve is really easy and interesting. You can have a working knowledge without getting your hand really dirty. You understand the game workflow and there is so much example, whenever you are lost you can find some explanation.

The main advantage of XNA over direct x is the inizialitation
It seems to be a silly stuff but believe me ,Direct x was a nightmare from this point of view
XNA use C# which is definetly simpler than C++
Beside XNA supplies a library of the most common functions

Having said that do not assume that XNA is a game engine
XNA is a framework or, if you prefer, a sort of " organizer " of your game \ graphic oriented code
Very important, do not undermine it

However you must code by yourself all the essential game engine features
An hard task believe me

In my opinion it can be of use to develop specific multi media applications \ simulations thanks to its flexibility

As far as games are concerned I would stick to a commercial game engine

If you are still interested I would suggest to buy the book
" XNA 2.0 Game programming Recipes " by Riemer Grootjans

If you put all the recipes togheter (abt 700 pages) you get a very basic game engine

Alberto..

which would you prefer?
A7?

As a game engine, you mean ?

I keep using A7 because it is stable, well documented and reasonably updated
These are the most important features in my opinion

However after using the XNA " intelisense " system I must confess that I am a litlle be disappointed about 3dgs

intellisense rocks, I agree it's one area that 3dgs could do with some help. Maybe it's possible to do something with eclipse or a possible plugin for V.S?

I'd write one but I'm not sure the returns would be worth the effort...

intellisense is a big lack of missing features.. but you are totally right... it has a great community, tons of docs and lots of features..

but I am not sure if it the right one for me..
I try to get to know A7, XNA, BLADE3D, Darkbasic (.net) and DX Studio..
Every engine has advantages and disadvantages.. and I didnt find the right solution for me...

Good physics, good coding and a browser pluging are things I need...

Same for me but A7 is the most balanced engine of the group

You cant really compare XNA to A7 in fact. If you want to compare A7 to another engine I would choose leadwerks, s2engine or shiva

Xna would be compare to the c# wrapper for ogre per example.

yes, you cant compare A7 and XNA..

but perhaps we can get some other impressions if we talk about it..
I personally dont like Shiva. Leadwerks doest not have good (C#)docs.

Thank you for some very interesting posts.

My main question was really: is A7 a serious game engine ? Personally, I love it. My critisism, bizarrely, is that it is too good. It just makes everything really easy. Surely there must be more pain ? Would an employer take me seriously ? I suppose this has been answered ( .." they would be more interested in your C++ skills ...") and so I shall shut up !

Happy New Year

Drac.

it is taken serious if a serious prototype is made and if you refer to it as a7 the engine, and your not using "3D Game Studio" the whole suite for your work, in fact. Its all about making the prototype impressive so that an employer will still trust in you skills. Although if you looking for a job, say in an actual company, unless you have some c++ skils or modeling skills, you really dont have much of a chance at much more than a qa guy or an intern.

I dunno man, I've seen some pretty shitty code when I worked as a game programmer. Probably some of the most untidy sloppy code I've EVER seen (and that was just mine)...

Seriously though. I agree on the C++ aspect, however, a half decent demo that is visually appeasing will definitely get your foot in the door. The 3d code you write in 3dgs isn't really much different than doing it in your own engine, just different function calls.

No, there is much of a difference. Just compare code written for 3dgs with code written for Irrlicht. Then you get a little hint. But if you compare to code written for Ogre the difference raises and last but not least the very well structured and object oriented code of the C4 engine is much of a difference.

The code looks completely different if you are using classes, templates, virtual objects, interpreted type conversions, singletons or other stuff that you will not find in a 3dgs code.

And there is a difference in C++, C# and Lite-C. The first 2 are supporting classes, inheritance and other modern techniques but you still have to care about pointers, references, memory leaks in C++.

You have to be a good software architect to create good, robust and reliable C++ code. It will be more easy to create good C# code even for a beginner.

And if you really want to create your own appealing demo then you might impress some friends but not a real game dev company. Such companies know about the real problems. Rendering and lighting a cube with a few OpenGL or DirectX commands are not really a problem. But collision-detection, scene-management, path-finding, dynamic shader creation on the fly, efficient shadow mapping, big terrains - these might be some impressive features.

If you come to an established company then they already use a technology. They probably need help to improve their scene-management, their shader library, their AI. They might be interested in converting a title to a game console. That is what you could offer them.

I have to support what ecg_limey wrote...
It didn´t was a big differnce between A7 and irrlicht when I tried it, other than irrlicht having some usefull features missing and also having some bad design (at least that is what I would call irrlichts way to handle shaders and those strange animators).
I think that ecg_limey is mainly talking about rendering things like water, sky, shadows, plants, ...
For which the concepts are always the same one may prefer to use a plugin for some of those things to have it really advanced but there is no big difference it is just a fluent process where you get to know more and more new stuff but you don´t really notice it as the concepts are basicly always the same.

Yes, as I wrote, Irrlicht is only a little difference. It is very intuitive and easy to understand. This changes with other engines.

And as I also mentioned: The real challenge is to code something more than just water, sky and plants. Slin, you already know this for sure. Your challenge is the shader / shadow route and I am sure this is really a knowledge that will help you later to get a job.

And most XNA developers know this as well. They have to code scene-management, physics and so on on their own.

There aren't any "BIG" different between 3DGS and Microsoft XNA, but 3DGS rulez!

Water with vertexanimation isn´t that easy as you need many vertices to have it good looking and thus some kind of LOD. You need some not that easy math to generate the animation and to calculate correct normals for it. You will need to find out the depth of the water and the amount of water between the camera and an object to get a nice shoreline and blending between reflection and refraction. Crysis for example doesn´t use the depth of the water but the amount of water between camera and an object for the shoreline. This value can be easily calculated in the refraction pass and can then be written into the rendertargets alphachannel. The disadvantage is that the shoreline will move if the camera moves depending on the terrains shape. If you are above shallow water with the camera and near to the surface and whan you are looking to the land the shore will be everywhere, but if you look from the land over the water and are very close to the watersurface there will be nearly no shoreline.
What I just wanted to express is that also water can be nearly as tricky as shadows... One could even create some water particles and calculate their flight direction depending on where and how they where created and the wind direction, one could also to animate the water depending on objects penetrating it and it is also a little tricky to make objects swimming on the surface to move depending on the wateranimation...

With sky I was talking about stuff like atmospheric scattering, dynamic (maybe volumetric) clouds and a shaderbased day night change.

Plants need to be animated, what about local wind sources like explosions? You need very good LOD and if possible every tree should look a little different. You will also have to do some good clipping and should organise the rendering very well. This is diffinatly not easy...

I doubt that many XNA developers are coding physics on their own as there are many very good free physicsengines around and to use those really isn´t a hard task.

True, but on the same token, does a company who gets hundreds of good apps a day really want to get someone who knows how to use c-script and thats it? Personally I'd like someone with even a basic understanding of c++ (not necessarily an expert) plus thorough knowledge of some other language (c-script, lite-c, c#, java, python, lua, etc.) that way they know a little of a lower level language if they need it, but to know they at least know how to script the game using whatever language the engine supports, and the company decides to go with.

seems you didn't understand well ! Just try to use xna ! You'll understand that you need to create your framework and a scene management by yourself, there is no short cut or whatever. Xna and 3Dgs are really different one is game engine the other one a sort of graphic library ! I mean just read the whole post before you start actiting as a 3dgs fan.

I did say foot in the door Of course, C++ is a requirement these days but knowing game studios, if you have 'something' that is outside of the average demo they will more often than not at least talk to you.

I've added intellisense into my application, 3DGS Easy Scripter. I knew that this had to be included because it makes life so much simpler.