Kyle Schouviller:

Then what's the point in including a networking library in 2.0 at all?

I guess if you don't use it, it would be like we didn't include it from your point of view. For the people who do want to use it, though, we couldn't not include it and have them pretend we did.

There is more to the big picture than we are revealing at this time. Unfortunately, not all the pieces can be put together at the same time. We definitely could choose to wait and deliver it all at once later, but we think it's better to deliver more incrementally so that you get a chance to see us making progress.

Like I said, we know you want more, and we intend to give you more. We just don't have it ready yet.

 

The XNA Folks have said a bunch of times that they want to create the YouTube of video games. Distribution problems are clearly high on the radar. One of the sad facts of life in being Microsoft is that you can't ever half-ass anything. Once you release something, you are comitted to it. You gain customers which you must support.

The "YouTube of video games" games vision strongly suggests big plans for distribution of XNA games. I have no idea what the current timeline or planned feature set looks like, but I'd imagine it makes a great deal of sense for the XNA team to not have to support additional distribution mechansism that will likely become obsolete in "vNext" whenever that is (3.0? 3.0 refresh? 4.0? 9.0? :-).

Networking is a particuarly funny bit because it depends on the Live servers. Live for Windows is brand spanking new. In the past, pros had to create seperate networking subsystems for their Windows and Xbox versions. Professional game developers who created networked games for Xbox can afford multiple Xboxes to test on, but indies generally can't. That is why the networking APIs must be available on Windows for game developers. Think of the networking APIs are Xbox only for now; the Windows support is for debugging only. You can create an alternative networking subsystem if you want to support both platforms for now.

One thing that I find truly facinating is that XNA makes game development so productive that people have created loads of great games faster than Microsoft can create a means to share them! Dream Build Play was expected to give out one contract, but four were awarded! That's poof enough to me that this success of the XNA community has been beyond expectations.

You've all seen the South Park episode featuring the Underpants Gnomes, right?

That's basically what is going on here.

Step 1: Add networking support.
Step 2: ???
Step 3: Profit!

The big difference between us and South Park is that we actually have a plan for step 2, and we have some of our little coding gnomes hard at work on it even as we speak. We just aren't ready to talk about the details in public.

I appreciate this can be frustrating, because if you are looking at step 1 and trying to figure out how you're going to get to step 3, there is obviously a piece missing. There really isn't anything more I can say here (without losing my job anyway!) other than to point out that you do have several options:

• You could trust that we know what we are doing, and that step 2 will get filled in before too long.

• Some people will find that what we are providing in step 1 is enough for them to do interesting work (actually I suspect most students and many hobbyists, who are after all our primary customers, will fall in that category) so they can use this productively without even caring about what's coming in step 2.

• You could ignore our steps 1 and 2, and find your own way to 3.
  

The key point, though, is that we have noticed the need for a step 2 and are on the case. I happen to feel that, even if it doesn't yet meet your particular needs, what we have in step 1 is still interesting enough to be worth shipping as soon as possible, rather than waiting and not shipping anything at all until every step is complete.

I'm sorry I can't talk more about the details, but I can't, so I'm going to stop typing now.

The ZMan:

Shaderboy:

Will all users be able to get access to the new networking API, or will we require a signed title with Microsoft?

The Networking API is available to everyone, however there are some extensions to the library to allow commercially signed games to get access to things like achievements which will require you so have a publishing agreement with Microsoft on both PC and/or 360.

One thing of interest that came up on the second day of the networking talk (i.e. NOT the one that was broadcast) is that the networking APIs will NOT be available in the XNA redist. So to run a networked game on 360 OR Windows using the XNA network library will require your end users to have Game Studio 2.0 installed (and therefore one of its dependencies like VS or C# express). In addition your end users will of course need an XBL gold account to play match making games and a creators club membership if they want to run on the 360.

I won't comment or speculate on why since I like being an MVP....

???
So how are we expected to make any money with this, at all?

Kyle Schouviller:

that leaves full VS integration as one of the few truly useful features of GS 2.0.

In the list of new features on the XNA Team Blog, there are a bunch of other features I'm dying for:

• Improved Visual Studio integration
• Processor parameters (no more Scale20PercentProcessor, Scale25PercentProcessor :-)
  
• XNA games hosted on a form (great for level editors!)
  
• Use the virtualized GraphicsDevice: no more special code to handle device reset and recreate! (This will make life so much easier)
  

I would like to congratulate the XNA team for all they have done.  I understand you have to crawl before you can walk, and since XNA is not even a year old yet, it is still finding it's legs.  I am more excited than anything to give this Networking bit a try, even if it is not yet mature.  Keep it coming boys ( and girls ).  I'm ready for XNA 6.0.

- Matt D

Shawn Hargreaves:

Step 1: Add networking support.
Step 2: ???
Step 3: Profit!

I love the analogy... very well put :)

Shawn Hargreaves:

• You could trust that we know what we are doing, and that step 2 will get filled in before too long.

• Some people will find that what we are providing in step 1 is enough for them to do interesting work (actually I suspect most students and many hobbyists, who are after all our primary customers, will fall in that category) so they can use this productively without even caring about what's coming in step 2.

• You could ignore our steps 1 and 2, and find your own way to 3.

Thanks for your reply Shawn. My only problem is not trusting you guys, I know that you've made awesome work in what, not even a year, which is unprecedented for a library of that quality. But I also know that the vision of developers are often not the same as the marketing guys.

So while I fully trust the XNA team, we cannot be sure of what Microsoft will decide concerning user generated content distribution, which is at the center of all GameFest news on the internet (regarding Bands of Bugs, Unreal, etc), and unfortunately not in a good way.

Anyway, I know that you can't reply to that without breaking a NDA, so I won't insist :)
I'll probably play with the network feature, because I have to admit that playing with profiles etc will be a lot of fun.

Righteous Tool:

In XNA Game Studio 2.0, we are still focused on enabling scenarios that no one else can do. Adding a networking API to the Xbox 360 is one of those things. One major utility of the API being there on Windows is that you can use Windows to develop and test your multiplayer games for Xbox 360. After all, every XNA Game Studio user has both a PC and an Xbox 360 console, but not everyone has two Xbox 360 consoles.

This helps me understand the decision. I was surprised when I first read this thread but thinking about it as a 'network API for XNA on Xbox 360' makes more sense. We are already limited with how we can distribute Xbox 360 games until 'step 2' is finished, so it's fine that the Xbox 360 network API is limited in the same way - in fact, thanks for making it work on my developer PC, so I can test without 2 Xbox 360 machines.

Cheers,
Leaf.

Don't get me wrong - I think that the networking stuff we've seen is absolutely amazing, and I can't wait to play around with it.  However, if I want to send a networked game I've made to my friend (who has a gold account) and say, "Hey, try this out with me," he's not going to be able to play it without installing GS 2.0, whether on his Xbox or his PC.  So then I start telling my friends, "go install GS 2.0, then get my game and play it."

Now someone is going to eventually come along and say, "everyone download this dll and put it in this directory - it's got the networking stuff from GS 2.0 so you don't have to install it all."  While that's definitely not legal, it'll happen, and for hobbyists who want to let other people play their games, I'd worry it'll happen before there's any official solution.  However, if these libraries were included with the XNA Redistributable (which seems to only be a non-technical issue at this point), then this wouldn't happen.

I know you can't tell us exactly why this isn't happening already, but if there's some way we can rally enough support to get it done, let us know.  I've opened the issue on Connect, and if there's anything else I could do, I'd love to know.

Shawn Hargreaves:

Some people will find that what we are providing in step 1 is enough for them to do interesting work (actually I suspect most students and many hobbyists, who are after all our primary customers, will fall in that category) so they can use this productively without even caring about what's coming in step 2.

I also had some comments about this.  What I saw at Gamefest was a huge push to make professionals aware that XNA can be used for professional game development.  This seems to be supported by the Live support coming with 2.0, and the extra libraries that are apparently available/nearly available for developers to add achievements and such if they've got all the stuff together and are making a professional title with XNA.

That's great, and I hope it all works out - it would definitely help XNA develop further.  However, where does this leave the "primary customers," who are seeing all these features come along that they can't make full use of unless they're developing a professional title?

Shawn Hargreaves:

The key point, though, is that we have noticed the need for a step 2 and are on the case. I happen to feel that, even if it doesn't yet meet your particular needs, what we have in step 1 is still interesting enough to be worth shipping as soon as possible, rather than waiting and not shipping anything at all until every step is complete.

I love what's happening with XNA, and I'm very impressed with how far it's progressed so far.  However, even without a step 2, it seems like step 1 is perfect for satisfying the needs we currently have (playing our games with friends and family) - except for the redistributable not including the same files that the development package has (even though both are free downloads).

JunkMailer:

This helps me understand the decision. I was surprised when I first read this thread but thinking about it as a 'network API for XNA on Xbox 360' makes more sense. We are already limited with how we can distribute Xbox 360 games until 'step 2' is finished, so it's fine that the Xbox 360 network API is limited in the same way

That's not a bad way to think about this: in fact the reasons networking is limited on Windows are actually very similar to why distribution of your games (regardless of whether or not they are networked) is limited on Xbox. In many ways, Games for Windows - LIVE is bringing a subset of the Xbox platform over to Windows (a limited and secure platform with a certification process, verified executable and user identities, and access to protected resources such as Live), so the challenges for us to expose those secured platforms to a wider audience are extremely similar.

In fact, it would be a very reasonable guess to assume the missing "step 2" involved in distributing Live games on Windows is probably not at all dissimilar to the missing "step 2" that currently prevents you from distributing your games at all on Xbox...