Armory's future is announced on github

[Sandy]

I think some people have not checked the github page, so I will share the information.(https://github.com/armory3d/armory/issues/1545)

[Didier]

Humm … Are we witnessing the all-too-common target drift that can kill a project? https://github.com/armory3d/armory/issues/1545#issuecomment-574267870

[olsen]

I hope Armory keeps on supporting Nodes tree editing for hobbyist like me in the future,goes on the spirit of BGE.

[Simonrazer]

Yeah, the Logic nodes are the primary reason I keep on coming back to Armory. They’re awesome and should be kept at all cost imo

[M_Ent8h]

Don’t worry Logic nodes will still be there. After all they power a lot of people’s work from what I saw, so not supporting them would be foolish.

[rogper]

Lol I’ve been waiting for the graphics to stabilize and now they will be completely re-made
It’s indeed a bit preoccupation… I’m still hoping to finish at least a level for the ACXT3R

As @Simonrazer mentioned, Armory biggest strength are the visual nodes even to me, they just feel like something that we (3D artists) can actually learn.

To me the best graphics engine that could be used would be Eevee, because any other means new texturing/shader methods that may look substantially different in Eevee/Cycles… I know that Eevee is not exactly a real time engine, but It’s a bunch of trouble having to make different shaders, specific to armory, that may not work in a short future… We need something solid, something compatible with the future to complete a project.

Indie guys may take 6 years to make a game, if the rules don’t get fixed then we can’t do more than demo scenes…

[BlackGoku36]

With C (Kinc), possibility of features increase greatly, Haxe did limit this possibility. For example, If you wanted to use Bullet engine, then you would have to generate bindings and do some more stuffs to make it work, but with C (kinc) we would not need to do that. This applies to almost everything that regards libraries, we will able to use other libraries such as Imgui, etc.
I think with C(Kinc), Blender viewport rendering will become easier task.

So, moving to C (Kinc) is +1, the only downside is the work that it will take to do it.

[Didier]

Sure @BlackGoku36 and we need a fine tuning of performances with easy connections too to existing low level libraries in C …

Thus too the next version of Armory in 2 years will be with KinAsm thanks to @RobDangerous who is doing a great job.

Be prepared to program shaders and logic nodes in asm and using something like that

GameLoop:
stdcall FillSurface,[Back],0ffh

ddcall DDS_GetDC,[Back],hdc
.if eax = 0
    invoke SetBkColor,[hdc],0ff0000h

    invoke SetTextColor,[hdc],0ffh

    mov [rect1.left],0
    mov [rect1.top],170
    mov [rect1.right],WSCREEN
    mov [rect1.bottom],230
    invoke DrawText,[hdc],MsgFullMode,0-1,rect1,DT_CENTER+DT_SINGLELINE+DT_BOTTOM

    ddcall DDS_ReleaseDC,[Back],[hdc]
.endif

jmp MessageLoop

It will be the best future solution that will push pc performance and memory to its limits using the asm and increase again the features …

You will have just to learn asm but then you will get a fine understanding of low level computing and an great inane amount of code optimization strategies ! good luck

Again be prepare too in 4 years to use a hightly parrallelisable asm usefull to exploit GPU capabilities extensively.

Are you ready ?

[M_Ent8h]

That’s a slipery slope If I’ve ever seen one. I don’t see the problem with wanting more control about what is done with for low-level parts of the engine, instead of having everything on a higher layer and hoping the compiler does its best.

The user experience should be relatively the same:

“Armory traits will be written in Haxe/JS like usual, or anything which compiles into WebAssembly”.

The only thing that may change probably is how the interface to those parts works, but it may even not change at all.

There might be a relatively big wall now for people who want to contribute on certain parts of the engine as it now it would require understanding C and knowing what you are doing. But then you look at Blender, and see that it’s possible.

[MikeRobinson]

Taking a bit broader view of this whole thing … “Quite obviously, someone is going to produce a next-generation, multi-platform game engine that leverages Blender” … but it certainly doesn’t have to turn out to be Armory3D.

So far as I can see now, the Armory3D project right now appears to be poised to suffer the same fate – notwithstanding(!) its then-perceived technical advantages such as they (were)(are) – of every previous open-source project that could not navigate its way beyond “its original founder’s thumb-on-the-scale” and proceed to the point where other people’s contributions actually made a meaningful difference.

“Oops… sorry… I ‘made a mistake’ by building Kha … or maybe by choosing Haxe.” But you see, “in the end what really matters is what I think …”

If Linus Torvalds had thought that way, in his dead-of-winter dorm room, we would not have Linux. Indeed, by a not-the-same stretch, we would also not have Blender. “Can you let it go?”