Microsoft Buys Github for $7.5 Billion

[psi29a]

Qt isn't the problem, OpenMW using Qt appears to be one of the occurring problems when compiling. If you look around on the github OpenMW page, it's fairly frequent.

If Qt isn't the problem, then help fix OpenMW's use of Qt.

Many developers also come here knowing DirectX and are not eager to delve into another graphics abstraction layer. So they leave again. Do you think your average Windows DX developer is going to start using OSG ? Learning + getting it to compile is it a time commitment which appears to be completely unnecessary when you already know something perfectly working, like DirectX. It's only interesting for Linux developers. I outright dare to proclaim; If OpenMW would use DirectX exclusively we would have OpenSkyrim already. I suspect this forum underestimated the number of game engine developers familiar only with DirectX.

DirectX isn't an option, it isn't cross-platform so that argument dead on arrival.

That being said, why does OSG have to do with helping to work on OpenMW?

There is more than enough work to go into OpenMW that doesn't have anything to do with OSG. So your bold claim is baseless. To further the point, when OpenMW did use DirectX, why didn't all the DirectX developers come flocking? Where is OpenOblivion and OpenSkyrim? You're just being disingenuous. Scrawl bent over backwards to make sure that OpenMW worked on OpenGL and DirectX renderers for years. Since he was the only one doing any work on the graphics side, he proposed that we switch to OSG then suddenly we have more Windows developers wanting to work on the graphics side of things. Why is that?

[psi29a]

This assumes I want to target multiple platforms...
Anyway, I will just continue de *nixing OpenMW upload to github and see if people jump on board for developing the Windows optimized version - and how fast we can proceed.

That's cool.

I hope to see future collaboration then.

[davidcernat]

Well I started working on it. It would be Windows only. It's simply a response to the harsh building environment on Windows, an evolutionary process adapting to the environment. One has to do the job, each dependency I bring down means less hassle to compile OpenMW for new Windows native developers.

Each dependency you bring down creates a massive code conflict with the main project and prevents you from taking in further improvements made to it.

Essentially, your ingenious plan is to split the project by creating a Windows-only OpenMW and disregarding all further improvements made to the official cross-platform OpenMW just because you have trouble following some build instructions for Windows.

It seems like entirely the wrong angle to me. Then again, perhaps I'm special: I'm a Windows developer who somehow managed to follow the instructions on his first try.

[AnyOldName3]

Hint: it's going to be exactly the speed that you on your own can proceed.

[Blue]

Well I started working on it. It would be Windows only. It's simply a response to the harsh building environment on Windows, an evolutionary process adapting to the environment. One has to do the job, each dependency I bring down means less hassle to compile OpenMW for new Windows native developers.

Each dependency you bring down creates a massive code conflict with the main project and prevents you from taking in further improvements made to it.

It may split into another direction, that's okay though. It's an open source project after all ?

@AnyOldName3

Hint: it's going to be exactly the speed that you on your own can proceed.

If some developers find it useful, they will catch up, especially after showing some candy on youtube (Windows mod compatibility and all, MWSE integration would be great). Or maybe nobody cares and I will just have my private Windows optimized version. It is just a fun project for me, nothing serious, neither is it the first time I did something like this.

[davidcernat]

It may split into another direction, that's okay though. It's an open source project after all ?

Of course no one is going to stop you, but we can still debate the wisdom of the direction.

If some developers find it useful, they will catch up, especially after showing some candy on youtube (Windows mod compatibility and all, MWSE integration would be great). Or maybe nobody cares and I will just have my private Windows optimized version. It is just a fun project for me, nothing serious, neither is it the first time I did something like this.

I can't figure out what you mean by "Windows mod compatibility." I haven't seen any difference between using mods on different platforms for OpenMW.

I also hope you realize that there are lots of real optimizations remaining to be made to the official project which your split will prevent you from taking in. It may just be the case that Windows builds of the cross-platform project end up a lot more optimized than your own fork.

Of course, I don't know you. It's possible you have a lot of experience with taking cross-platform projects, splitting them up and providing the definitive Windows experience for them. Yes, it takes serious effort and dedication to successfully run a project like this day-in and day-out for years, and not any random completely unknown person popping out from nowhere can pull it off... but I don't want to prejudge. Perhaps you can show me some of your past successes?

[Blue]

I can't figure out what you mean by "Windows mod compatibility." I haven't seen any difference between using mods on different platforms for OpenMW.

There are mods for Vanilla MW not running on OpenMW, due to missing binary compatibility, that's what I am addressing here.

I also hope you realize that there are lots of real optimizations remaining to be made to the official project which your split will prevent you from taking in. It may just be the case that Windows builds of the cross-platform project end up a lot more optimized than your own fork.

I know the Windows NT family fairly well and I certainly know some functions which can be made a lot more effective than they are now by using the native API. Starting from the filesystem up to the rendering engine using DX12 low level hardware access. Not saying I will do it. Doing anything to replace OSG is a time commitment I am not willing to take. It's just some fun playing around with OpenMW, shouldn't be a problem I assume ;- D
What people here don't like is to make things Windows exclusive, that's easy to recognize. I more or less expected heavy resistance.

Perhaps you can show me some of your past successes?

No, it's on proprietary software, so I can't really provide insides on this forum. Was it a success in the long term ? Debatable, but I did once got it through.

[AnyOldName3]

There are mods for Vanilla MW not running on OpenMW, due to missing binary compatibility, that's what I am addressing here.

This kind of mod will literally never be compatible with OpenMW, no matter how much of it you change. If you make a mod that modifies the original engine's executable it's not even slightly going to work when it tries to apply its modifications to a different executable.

[Blue]

There are mods for Vanilla MW not running on OpenMW, due to missing binary compatibility, that's what I am addressing here.

This kind of mod will literally never be compatible with OpenMW, no matter how much of it you change. If you make a mod that modifies the original engine's executable it's not even slightly going to work when it tries to apply its modifications to a different executable.

Yeah, some rely on specific hardcoded addresses. Making it compatible is NOT impossible just really inconvenient compiler work. Far more practical would be the integration of MWSE. I don't think any mod author would change their mwse mod for compatibility to fit the past 1.0 OpenMW scripting extensions. Why bother with compatibility when you can make your own interface ? Apple ftw. Always a great idea.

[davidcernat]

There are mods for Vanilla MW not running on OpenMW, due to missing binary compatibility, that's what I am addressing here.

If you're talking about lack of MWSE compatibility in OpenMW, I don't see why that needs to be addressed by a Windows-only fork. You could just have a cross-platform fork with MWSE compatibility and save yourself a lot of trouble. Your YouTube videos would then also avoid mixing up unrelated project scopes.

I know the Windows NT family fairly well and I certainly know some functions which can be made a lot more effective than they are now by using the native API. Starting from the filesystem up to the rendering engine using DX12 low level hardware access. Not saying I will do it. Doing anything to replace OSG is a time commitment I am not willing to take. It's just some fun playing around with OpenMW, shouldn't be a problem I assume ;- D
What people here don't like is to make things Windows exclusive, that's easy to recognize. I more or less expected heavy resistance.

Pros:
- You gain a marginal performance improvement (possibly unnoticeable)
- Windows developers theoretically have an easier time building the project

Cons
- You throw away improvements done to the official project in the future, including significant performance improvements, countless bug fixes, new features, etc.
- You prevent Linux and Mac developers from contributing
- You try to take developer attention away from the official fork where all the real work is done

Call it heavy resistance. I'll call it common sense.

I can spare 1-2 hours at evening removing libraries like OpenAL, Qt, Boost over time.

I outright dare to proclaim; If OpenMW would use DirectX exclusively we would have OpenSkyrim already.

Perhaps you can show me some of your past successes?

No, it's on proprietary software, so I can't really provide insides on this forum. Was it a success in the long term ? Debatable, but I did once got it through.

Unless you're too far on the wrong side of the Dunning-Kruger effect, you need to take a step back and ask yourself if – as an unproven quantity even to yourself – you really know what you're talking about by suggesting we'll move closer to "OpenSkyrim already" with your "1-2 hours at evening" dedicated to killing cross-platform development.