[Frank]

I've been interested in understanding more about the planned switch to it.

From what i could understand, it seems like that update would be necessary in order to eventually move away from XNA (which is or will be obsolete relatively soon).

As an user, though, i'd be interested in the pratical, tangible results of that preliminary move...so, what the move to XNA4 will mean?

[Genma Saotome]

IMO it would be better to put the effort into either Monogame or SharpDX, both of which could be thought of as as something akin to .XNA 5 (if it had been produced).

The real problem here is XNA is dead and it's not obvious to anyone that a jump to either of the free versions of the commercial game engines could allow continued use of MSTS routes and content, nor that the OR team has the resources and willingness to take on such a major effort.

[James Ross]

Frank, on 09 November 2015 - 04:37 PM, said:

As an user, though, i'd be interested in the pratical, tangible results of that preliminary move...so, what the move to XNA4 will mean?

The only thing that comes to mind is that we may be able to produce a 64bit version (at least with MonoGame/SharpDX, probably not XNA 4 itself) which would be nice, if not essential, for some people.

ISTR that when I looked a little while ago at XNA 4, there was actually something which would be a serious problem (might've been dynamically compiling shaders). If I'm remembering right, we should probably do what Dave says and move directly towards the post-XNA options.

[Frank]

So, shaders would need to be reworked if that transition is implemented.
Might that (at a later time) mean the possibility for further improvement on them?

What's there now is amazing, and has little resemblance with old MSTS, but someday i'd wish to see no difference between powray and OR :).

[James Ross]

Frank, on 10 November 2015 - 08:24 AM, said:

So, shaders would need to be reworked if that transition is implemented.
Might that (at a later time) mean the possibility for further improvement on them?

If we skip XNA 4 we probably don't need to rework them, and I'd say any improvements to them would be orthogonal to this - we're not actually near the limits of the current shader model (3) IIRC.

[Frank]

Just for curiosity, what else can be implemented with the current model, which isn't already there?

Plus, it doesn't look like the switch to 4.0 might be the best decision, yet it's listed among the steps towards 1.1..

[James Ross]

Frank, on 10 November 2015 - 10:32 AM, said:

Just for curiosity, what else can be implemented with the current model, which isn't already there?

I don't think we can do deferred rendering "properly" with DirectX 9, but other things we can do include bump mapping, environmental mapping, a variety of multi-texture effects (including global illumination IIRC).

Frank, on 10 November 2015 - 10:32 AM, said:

Plus, it doesn't look like the switch to 4.0 might be the best decision, yet it's listed among the steps towards 1.1..

Well, the "Upgrade to XNA 4.0" item may come off or be replaced by something else; it originally looked like a good move.

[Frank]

James Ross, on 10 November 2015 - 11:37 AM, said:

I don't think we can do deferred rendering "properly" with DirectX 9, but other things we can do include bump mapping, environmental mapping, a variety of multi-texture effects (including global illumination IIRC).

Those would be fantastic features, once it will actually be possible to use them (i'm thinking about current .s file format, for istance) :) - glad to know the potential is there.

[SP 0-6-0]

Well, If a switch is made to a open source XNA replacement then maybe ORTS will get a little closer to having Linux support which will finally deliver a decent train simulator the Linux community.

I don't know much about which version of XNA ORTS is using to decipher what XNA4 would offer vs the current XNA framework?

Robert

[James Ross]

SP 0-6-0, on 12 November 2015 - 08:14 PM, said:

Well, If a switch is made to a open source XNA replacement then maybe ORTS will get a little closer to having Linux support which will finally deliver a decent train simulator the Linux community.

Perhaps. If we used MonoGame, we'd get some Linux support from it, but if we used SharpDX, we wouldn't. Since MonoGame uses SharpDX on Windows anyway, that choice would depend on whether the MonoGame API provides the necessary means for us to work.

I did try running MOMA, Mono's tool for detecting compatibility with their framework, in an attempt at seeing what (beyond the obvious) we'd need to change to work on Linux, but it was a catastrophically bad tool and impossible to run by automation. :(

SP 0-6-0, on 12 November 2015 - 08:14 PM, said:

I don't know much about which version of XNA ORTS is using to decipher what XNA4 would offer vs the current XNA framework?

We're using XNA 3.1.