rpg.  no doubt.  or just a simple character wandering around in a single simple level, doing whatever.

Actually I would love to see something like that too, however devil may cry, onimusha may be a bit too much, starter kits are supposed to be simple and easy to understand.

However a simple hack and slash game (specially one in 3d, since is really too easy in 2d) could be a great starter kit for all kind of projects. RPG, Adventure, Beat em up, even fighting games.

All we need is a character walking in an eternal plain and beating another single character (a giant ant per example) if the guy has two attacks (fast, strong) that would be enough code to add any number of attacks. and with just 2 characters (minion, boss) we could have a large number of enemies.

Somthing I think would be cool as I see it as the one of the future technologies in games is synthesized character animation. It would be cool to have a sample that covers that. However, I don't think it would end up being feasible as I believe it is a far larger subject to cover then I can fully grasp.

For those unsure what synthesized movement is. You will see in the new GTA game. To my knowledge they are not using any caned animations.

Vangald:

Somthing I think would be cool as I see it as the one of the future technologies in games is synthesized character animation. It would be cool to have a sample that covers that. However, I don't think it would end up being feasible as I believe it is a far larger subject to cover then I can fully grasp.

For those unsure what synthesized movement is. You will see in the new GTA game. To my knowledge they are not using any caned animations.

That would be better covered in a sample or article. It wouldn't make much of a starter kit, since the idea of a starter kit is to give a starting point for a game.

 

Ah. Yeah. Good point.

Any idea how far the old style rpg starter kit is progressed? If it has been started anyway ;)

It's looking good. We got to see it in action at the MVP Summit. It'll be interesting to see what people do with it.

Wow, great thread.  I'd like to summarize some of the popular requests / comments:

• Starter kits should be simpler in architecture (as relative to the "thing" they are starting)
• Demonstrate core concept(s)
• Easy to build on and extend
• Perhaps not a "complete" game?  That is, maybe not provide all the robustness code that a complete game needs? (Menus, profile checking, signing in and out, storage, options, save games, pause menu, etc).
• A more functional project template would be good (but different than a starter kit)
• Each starter kit should focus on a particular game style
• Game styles requested
• RPG (coming soon!)
• 2D Platform (Lode Runner, Mario, Metroid, etc)
• 3D "shooter" (either first person or third)
• RTS-ish
• Animation is still a topic that needs to be better explored.

Anything else?

Starter Kits should use the sample code you provide in the education section. Once people understand (for example) the GameState sample then they know how it works in ALL the starter kits instead of having to work it out. Same for particles or instancing etc.

 

So should all samples have an eye towards how they can be used together or in a starter kit?  Will this lead to feedback on the samples becoming too cluttered (as it will most likely take some amount of code that isn't directly related to the topic the sample is focused on).

Is there another category of materials that are closer to "components" (be they GameComponent or not) that are directly intended to be building blocks for a game?  I'm not sure I agree that samples should be building blocks, although a successful sample should allow the user to implement the concept in their game.

 

 

If a starter kit is purely an example of waht can be done then do whatever each author needs.

If they are supposed to be learning tools then the gap between the example code and the kits is too big today. If segments of each game were consistent then it would go a long way to bridging that gap. Using bits from the other samples would be a good way of getting some consistency. I don't want everything included if its not needed but all games need some menus and game state so if that all looked the same or similar we could basically learn it once. Right now you have to learn every line of every kit to get anywhere significant.

I'd like to see some of the starter kits focus more on game style instead of common programming tasks (such as menus, etc.). It'll be great if samples such as the GSM were wrapped up into an application building blocks and just referenced from the other samples instead of including the GSM code in the other samples, that way the sample's code doesn't get overwhelming when it includes the same code as other samples. Instead of including all of the GSM code, just use a GSM library and reference those libraries instead.

By turning things like GSM into application building blocks, it'll be easier to integrate into other games (while still giving the source code for those of us that want to tweak it). That also helps keep everything consistant across the kits and samples and keeps the focus of the samples on a particular topic instead of adding in all of the code of all of the other samples.

Has already been requested a few times, but a simple 1st person or 3rd person shooter, as it'd be nice to have some decent assets for that type of game. The skinned model from the skinned animation sample is awesome, but as far as I'm aware, it only has a walking animation.

Being able to switch between a 1st person and 3rd person viewpoint, as in Gears of War, would be a bonus, but not a requirement.

It'd also make a good starting point for someone who wants to experiment with implementing AI for that type of game.

Something that Mitch Walker and Jim W touched on is the portability and reusability of code. XNA has gone a long way in embracing these concepts with GameComponents and Services, but very few of the samples actually prescribe to this model.

I agree with Jim, things like the base objects from the GSM should be bundled into a library where it makes sense. (ScreenManager, GameScreen, etc) I've done this myself, and its quite usable. Then, new developers could simply reference the GSM library and build on it from there.

Similarly, with other samples, if they were written as stand-alone GameComponents, then new users could simply drop them into their projects and go. A good example would be the ChaseCam sample - having a ChaseCam or FirstPersonCam game component / service would help a lot of new developers get off the ground quickly.

Changing the subject slightly, I'm also a little worried about the samples and starter kits being thrown together haphazardly. While a particular sample may be about one topic, the rest of the sample still gets looked at for examples of how to do things. The author may have simply hacked together a quick workaround to get their sample working, and new developers learn this as 'the right way'. I'm sure you see the problem here. Also, if starter kits are meant to be 'base' from which developers build their games, surely they should be written extremely well and optimised. You dont want someone's game to get poor performance because of something thats not quite right in the base.

I am in no way trying to insult any of the developers of the starter kits or samples - you guys have done a great job so far and all your hard work is very much appreciated.

someone knows when the rpg starter will be out and if it will be in 3d  or 2d ?

cheers

A complete game, though not very "starter", would been a very interesting study, since it would introduce main design flow patterns, as well as get into technical stuff about how certain graphical effects can be accomplished.

How about an RTS starter kit. That would rock!

I think that a game state management starter kit is a good idea, but at the same time I don't think it's a big deal - GSM is fairly simple programming when compared to game logic and rendering. I quite like doing it: it gives me a chance to run away from the complicated stuff for a few days. But I understand that others probably don't feel this way.

Here's a fun suggestion: get a copy of Macromedia (now Adobe) Flash, and make your entire menu in it. The simple motion and artist orientated drawing means that you can make something that looks pretty good, and then can port it to XNA. Plus it means that you know exactly what you're trying to do before you even start programming in XNA.

I think a 2 player side-on fighting game (Tekken style), with a few fairly simple animated combos, would make an excellent starter kit. I'd rather see it in 3D than 2D, though: I'd really like to see some examples of importing and changing between animations. Just my 2 cents.

EDIT: Whoops, sorry guys... I only read the first page of posts. If this post seems a little disjointed, that's why.

I'd like a good old fashioned Street Fighter type 2D fighting game starter kit. That seems like it would have the most unique content (The only 2D games that have been released so far have been tiny top-down affairs with simple collision).

Cardinal4:

A complete game, though not very "starter", would been a very interesting study, since it would introduce main design flow patterns, as well as get into technical stuff about how certain graphical effects can be accomplished.

 

Some of our earlier starter kits, like Ship Game, may be useful for this. 

 

You can also check out our mini games, which are complete games, released under the Microsoft Permissive License.  The upcoming "Robot Game" will be one of these.

 

You may also find some of our samples useful for investigating particular graphical effects.  For example, here's a list of our educational items that relate to graphical postprocessing.

My vote would be for platform starter kit. It needs to demonstrate:

1. Level and sloped platforms and how to make the player walk along these (assign a slope to a tile?).

2. Moving platforms that the player can "attach" to, to be carried somewhere. Once you've got this you've paved the way for switches activating platforms under code control.

3. Ladders of course, and perhaps other climbables.

4. Collectibles and how to manage them (quadtrees or similar).

 5. Save and restore state.

6. Ideally, how to integrate some flashy 3D effects into the 2D world eg lighting for fires or lava.

The RPG kit went a huge way further than I was expecting - you could carry over the portal concepts and tile engine (possibly) into a platformer starter kit to give you end of level/load new level functionality. Integrate the GameStateManagement example to get the front end.

Anyway, I think MS are doing a great job. I only wish I had the brains for 3D.

It ultimately depends on what you want to teach your users.  I would think that you'd want to teach them concepts so I totally agree with Mitch's comment:

•  Starter kits should be simpler in architecture (as relative to the "thing" they are starting)

 

Thinking about this sort of question reminds me of the days when people were trying to learn, say OpenGL, and NeHe's tutorials became pretty popular as they taught concepts not complete games.

Plus it seems like a smaller time commitment to teach this way from the teacher's perspective unless you've really taken care of your code and re-using it is very simple (UI, input response logic, and etc.).

I would really like to see a 2d/3d street fighter style game or Mortal kombat style game. I have some great idea but need some where to start.

A 2d/3d style fighter would be good to see how to start it off. but i think that a action adventure might by good also because there could be model collision for the character to attack and move around in a 3d world instead of always shooting. not to say that FPS are bad but sometimes the old button bash gets that itch that fps don't get sometimes

Devinmccloud:

I would really like to see a 2d/3d street fighter style game or Mortal kombat style game. I have some great idea but need some where to start.

 

Me too!  A Fighting Game Starter Kit would be awesome!   I really hope to see one soon!