[stationmaster]

Hi everyone

I have not been working with XNA before, but I am trying to learn since ORTS is built on it.

However, I discovered that Microsoft no longer supports XNA. Now ORTS works on Windows 8, but what will happen in future releases of Windows?

What are your thoughts on this? Any ideas for migration? If so, to what?




BR,

/J

[Genma Saotome]

The OR team is aware of the issue and various alternatives have been looked at and discussed. AFAIK no firm plans yet. Options range between "Don't worry about it for a long time" to "Here's an interesting replacement game engine" to "Just use native direct X" and most anything else as well.

I'm sure some of the developers can explain the issues in great detail and clarity and perhaps a couple of them will.

[James Ross]

 stationmaster, on 06 October 2013 - 03:01 AM, said:

However, I discovered that Microsoft no longer supports XNA. Now ORTS works on Windows 8, but what will happen in future releases of Windows?

What are your thoughts on this? Any ideas for migration? If so, to what?

It is my belief that, as long as DirectX 9 and .NET exist on the desktop, we'll be fine with the current XNA 3.1 platform. That should be quite a few Windows versions yet.

We're not going to keep using XNA until we can't, however. There has been experimental work to use XNA 4.0 (different API to 3.1, and one that's supported by a number of non-XNA platforms), ANX, DirectX through SharpDX and SlimDX and even MonoGame to name just a few.

It is worth noting that XNA, as we're using it, does not actually provide very much functionality - it is almost exclusively a thin wrapper on DirectX (similar to SharpDX/SlimDX). That is positive, because it means we're only really tied to DirectX, and if that disappears there'll be more problems than just Open Rails. :lol2:

Just recently, in X.1733, we got rid of the XNA Content Project, which tied up the development environment to XNA (Visual Studio/Express 2008) so we're now pretty much free to use any version of Visual Studio (still a few minor kinks).

Replacing XNA is naturally a big project, but we've got options - quite a few of them. :)

[gpz]

I would like to make a note: I think, we use XNA not just as a thin wrapper on DirectX, we also use its MathHelper library intensively, that is responsible for a huge part of the backgrounding matrix mathematics. So anyway, for a move to a more native 3D library, a replacement mathematics library would be needed to be found as well. Knowing this I would like to add one more possible porting target to the list: OpenTK. That one is a C# wrapper library for OpenGL with a more or less XNA compatible math library included.

If it is possible to make a vote for the future, than there is one thing that I wouldn't like to see to happen in any case: a change that makes native linux port impossible. So in this regard a move to SlimDX, SharpDX, native DirectX would be really bad news to me, while a move towards e.g. MonoGame is right (that also contains the corresponding math library). However once I made a 95% port to XNA 4 in the past (that is a must for moving towards MonoGame) with not much interest of the other developers, so I assume this move is unlikely.

[James Ross]

 gpz, on 07 October 2013 - 01:25 AM, said:

...we also use its MathHelper library intensively, that is responsible for a huge part of the backgrounding matrix mathematics. So anyway, for a move to a more native 3D library, a replacement mathematics library would be needed to be found as well. Knowing this I would like to add one more possible porting target to the list: OpenTK. That one is a C# wrapper library for OpenGL with a more or less XNA compatible math library included.

I don't think we use any of the really interesting parts of the maths library, and could replace it with minimal effort IMHO.

And of course, changing from a DirectX wrapper to an OpenGL wrapper would be huge amounts of work and degrade performance.

 gpz, on 07 October 2013 - 01:25 AM, said:

If it is possible to make a vote for the future, than there is one thing that I wouldn't like to see to happen in any case: a change that makes native linux port impossible.

Whilst this is an admirable aim, I will not be discarding options simply because they make a Linux port impossible (or effectively so). An XNA-derived API may be a good compromise in platform support in the end, but if it turns out that e.g. SlimDX would be a decent performance boost, I'd strongly like to take it.

[rdamurphy]

The number of Linux users worldwide is between 1 and 1.5% Not really worth creating a version of OR - or any game - for.

Robert

[Genma Saotome]

IMO it would make more sense to be compatible w/ game consols than Linux.

Has anyone looked into what is required to get a free game into Steam?

[disc]

Linux compatibility means compatible with every platforms including mac. Windows compatibility means not runs on anything except windows ;)

[Lindsayts]

 disc, on 21 October 2013 - 08:44 AM, said:

Linux compatibility means compatible with every platforms including mac. Windows compatibility means not runs on anything except windows :p

The tone of this post may upset a lot of Elvastower people, you must understand that some of us have a different outlook on life and do believe in the Linux ideals.

Sadly Disc I believe your wasting your time. I get the strong impression the OR community is hostile to Linux. This would in fact go with the opinons of most Windows users, who are generally philosphicaly opposed to the Linux way of life, ie cooperative development.

It's probably of little use to point out such projects as...........

http://www.mono-project.com/Main_Page

a project to allow simple cross platform development and running of .NET programs.

and

http://monogame.codeplex.com/

another project to allow cross platform development and running of XNA games, mind you they are using 4.0 as there base.

I have seriously considered for somtime now of dumping MSTS due largely to the great hostility one runs into when you try and modify some else's work. A large amount of MSTS's material will dissapear due the style of licence most such material has attached to it. I consider it a great tragedy that works of people such as Alan Lownsbourough and a lot of people on this forum will eventually dissapear due to the phisophical views held by the developers.

Lindsay

[copperpen]

Personally, I am not bothered what it runs on as long as it runs. There is however a sizable msts user base and I hazard a guess that not many of them run Linux as well. I am firmly in the camp of " if it runs and works for me, what is the point of changing my operating system". Windows has always worked for me ever since the days of Win3.0.

The problem with dumping all of your MSTS content is that there will not be much left to run in OR until the facility to make routes and stock is available and file formats are agreed. Far as hostility goes, I do not have that problem because I modify whatever I like and do not distribute it.