BGE is dead. And the slow death of blender

Game Engine Game Engine Support and Discussion
105

Hi, I said 2020 randomly (in another thread) (Some time ago, Ton was talking about 2.8.5).

If I said that, it is because I follow 2.8 development ~every day in my branch (the one where I plugged logic bricks on eevee render).

The work they are doing is pharaonic. The Blender codebase is really really enormous and it seems they are rewriting a big part of it. I think we have to wait that 2.8 new viewport is stable enough before we can expect works on “interactive mode” (I don’t know why I had a “bad” feeling just because of this name but not anymore now) to begin. They are doing a great work but this will need a long time (they begun in 2016 and the viewport is far from being ready).

I guess BF devs are already under pressure and there is no need to hurry to begin to work on the new engine.

Beside that, there are important design decisions to take and (I think) a complex context to take into account.

So I guess it’s better for Blender, Blender developers, other context actors, to take the time it needs to choose a good design for the future interactive engine.

At least, interactive mode first scopes are already defined (physics, export as runtime, python API, visual logic), and as everything will be rewritted from scratch, this might take some time. But this is great as first scopes! And this will need hard work from Blender devs.

Furhtermore, we know that some of Blender devs are interested in “interactive” topics, so this is great. When viewport will be more stable, the interactive project launched, and with time, I really hope that this new engine will bring back BGE users to Blender.

Lordloki said that he wanted to contribute. I’d like to contribute too (if I think I can be of any help on a topic which interests me (maybe work on access to some bullet functions with python for example)). I hope too old bge devs will be interested too (I hope Tristan too one day will be interested!).

2.8 project is really amazing! It will just need a long time.

I was really skeptical/pessimistic/angry during a very long time but I’m more enthusiast now and I bring my support to BF and Blender devs who are doing an awesome work :slight_smile: No matter the time it will take, I’m more confident towards BF, devs and Ton :slight_smile:

106

I don’t know what this statement is trying to prove, so if an idea was not acted on immediately then it won’t happen?

There was a span of many months between what were just ideas for 2.8 targets and when we actually saw something that was functional and visible.

Another question for those who plan to stick with UPBGE and promote, you’re going to have to consider this question if you are to build a userbase. Would they rather use an engine that has a wide range of built-in tools under development that are visual, based on fast C++ code, and can be implemented with little setup and code, or would they rather use an engine where many of the same tools are hacked in systems using hundreds of lines of Python and requires a bit of effort and setup to get into their own game in a working state.

I know some of the die-hard UPBGE fans would not mind the latter simply due to the BGE’s workflow, but many users will see it as a dealbreaker (I remember the thread about a working dynamic ragdoll concept in the BGE, the Python-based setup was crazy hard to do anything with if you wanted to use your own mesh).

107

@Ace_Dragon @aWeirdOwl You guys are forgetting the main concern here:

And how if what hanzo says is true, make you believe, that Ton or BF will put such a hard work into “interactive mode”.

1 Like
108

Interactive mode is a very old topic and Ton expressed his wish to replace bge with it since a while. It will be a big challenge and early versions will surely be lite. But 1) it is an engagement to do it and there is already a task assigned to Campbell Barton to do it. 2) it’s good for Blender 3) most of interactive/game topics are interesting and fun to code 4) there will surely be interested volunteers…

109

Thank you hanzo. It is good to know the vision of the same conflict from all parts involved.

It is good to take decisions…

1 Like
110

Yo frankie did next to nothing to further the game engine…and it was clear to any game dev that there was no serious backing for it from the the BF…

However, I do not know Ton…and will not put my foot in my mouth over all this…I do find hanzo trustworthy from my experience.(from the forums)…

on a side note, I merely cam here because I saw loki changed his avatar…and I must say I’s awesome…the only real reason I came into this thread :rofl:

111

(I have nothing against Hanzo (and I know he’s an important member of BGE community and I played his BGMC entries when I contributed to upbge), neither Ton and I don’t want to enter their conflict). I just think that interactive mode can become great for many people (finally ~understood what meant integration with depsgraph and it can offer a big power to Blender), but we have to be very patient because all will be rewrited from scratch (though some parts of BGE code could surely be reused/adapted)

112

The difference is that the Interactive Mode will not only have much better code, but it will also be even more tightly integrated than the BGE was (actually using Blender code as opposed to its own implementation for everything). When you look at it, the BGE had its own code for just about everything and shared little with Blender. This could be seen when the user pressed ‘P’ and has to wait for all of the data to be converted, that is if the BGE does not discard some of the data because it does not support it.

As I mentioned before, 2.8 concepts like Workbench were only ideas until the last month saw serious development. Also keep in mind that the BF can’t work on every 2.8 concept at once because of resource constraints.

Also, it was only expected that the BGE community would engage in some pushback when the engine was removed, but threatening to destroy Ton’s livelyhood (and disrupt the Blender project by extension) would only cause the rest of the community to compare BGE users to toddlers who can’t handle not getting their way.

However, I’m pretty sure that Ton and co. would welcome the former BGE community’s assistance in getting the Interactive Engine to a state the BGE wished it could be, that is if people here make the decision to try to become proactive for Blender rather than trying to destroy it. What would not be liked when users get the assurance that whatever they have in the viewport will work in the game they are working on (without having to worry as to which features can and can’t be used). What would be wrong with a full WYSIWYG workflow as opposed to the current one that does not always work (because you used either an unsupported feature or the effect requires tons of python trial and error).

113

Not? that project has given us at least GLSL and the option to use normal maps.
i find this more then nothing.

114

And that’s a deal breaker for me. From what I saw in youtube - Ton was talking, how animators will be supported with the “interactive mode” and will make their little animations more “interactive” and that was it. Please prove me wrong with videos of Ton saying otherwise.
Then on gidhub Cambell (who actually is a great guy, at least that’s how I remember him from 2007-08, when he was active on the bge forums) was acting somehow distant from the idea for a complete game engine with standalone capabilities…he was totally unaware for what the community wants in the “interactive mode” …just how Hanzo said…Tons’ “brainwashing magic”.
At least he promised that “interactive mode” will have a standalone export…but any programmer would know, that this will require a standalone engine - aka “blenderplayer.exe”, and there will not be “blenderplayer.exe” in 2.8.
So enough BS.

P.S. The only way to support BGE in Blender is how they’ve done it until now, by duplicating the code from Blender in the engine. And that’s perfectly fine…the idea of integrated engine in Blender is great and it requires duplicating the code, which requires money and people…clearly Ton doesn’t want to redirect money to BGE.
He was direct enough in the interview - it cost a lot of work and we will change it with something in-between that will work as “engine” but just for animators…aka “interactive mode”. (Not a quote, just what he ment actually)

3 Likes
115

Can we please wait until we actually see something instead of having the assumption of everything set in stone (much like how current 2.8 development is smashing a lot of concerns people had previously)?

116

Right, nothing is set in stone… there isn’t even as much as a document outlining planned features. I’d really like to have at least an idea of what it is that we’re supposed to just patiently wait for. All I’ve seen until now is devs saying they need to stabilize 2.8 first and that interactive mode itself is “still under debate”, which doesn’t sound all that assuring to me.

I don’t like you.

1 Like
117

Heard Ton say that game engine will be a «crucial» topic. Heard too talking about a «better game engine». But the viewport needs to be finished and stable first. I wouldn’t have give 3000€ if I wasn’t sure we’ll be abled one day to do «games» with interactive mode… Anyway all of this is opensources and coders do what they want with… and it is already possible to do games in my branch. But I don’t want to continue the work in the direction I took (even if it is perfectly doable) because they’ll do better than what i can propose with another design

118

I can’t wait for 2.8 ige (interactive game engine)

but for now there are unresolved issues we can solve for bge.

we can port a forward+ pbr render to the GLSL varient of the pbr shader coder, and not touch any old material code/ rendering

this adds redundancy but also avoids old code, and allows for flexiblity
(NPR and PBR materials on 1 object using masks for instance)

with forward+ and PBR the old BGE/Ketsji engine will still likely be in use for decades. after the 2.8 project finishes, the old code can be rebased on 2.8 and the glsl version can be bumped
(after the last of the old immediate mode drawcalls are removed)

I think 2.8 has room for upbge/ketsji and IGE

1 Like
119

We all know there is room for the BGE and Interactive mode . That they deleted it instead, without consulting any BGE coder to help give it the minimal work needed to run in 2.8 (it DID run in 2.8 mostly until it got deleted) speaks more than Ton, acedragons or anyone elses words can.

Bold claims, but they’re just claims, forgive us if we’re skeptical that someone can come along and code something better. Maybe the BGE community would give you less blowback it… i don’t know, maybe you supplied an actual alternative offering before pulling out the part of blender they loved?

120

I don’t see how they’re any bolder than the claims the community here makes about the BGE (most notably the idea that 19 massive Unreal 4 releases and a decade of rapid Unity development cannot even come close to the functionality of the slowly developed BGE and later UPBGE). It’s also the same claim against Godot (despite the hundreds of developers outside of the core team who contributed via thousands of patches).

Also, it shouldn’t be that hard to write a new game engine with better code if the one being replaced has historically been filled with hacks and bandaids (the UPBGE cleaned everything up to an extent, but comments in the thread following it suggested that sizable bugs in basic functionality still came up and persisted for multiple releases). It didn’t take long for Eevee to become what BI wished it could be despite its use of rasterization technology (because the code was much better quality and is much more attractive to developers).

121

Considering that EEVEE has been in development since before December 2016
https ://code.blender.org/2016/12/viewport-project-plan-of-action/

2 years is a long time to produce something that is still incompatible with huge swarths of existing code and still crashes while doing nearly anything.

So no, I’m not optimistic that the BF can deliver a better game engine at this point. I’m thrilled if they delivered on their claims honestly but at the same time im not holding my breath until I see something more then a few blog or interview mentions and 20 minutes worth of vague design documentation.

Interesting side note while I was researching the development history of eevee:

10%(52 non BGE to 8 BGE posts) of the comments in the link above were asking about the BGE, thats 1 in 10 following the development closely were excited/worried/curious about how this would benifit the game engine.

Months later that number of BGE related questions increased to 13%(58 non BGE to 8 BGE posts)
https ://code.blender.org/2017/03/eevee-roadmap/

Sometimes I wonder if us BGE users are being painted as a few marginal, over dramatic grumps so its easier to dismiss us. Tell us to do nothing and wait patiently until the community evaporates and then no need to deliver anything past the initial for animators stuff.

122

What needs to be noted is that it was several months after the announcement was made that Eevee development got a full-time developer and really got going, so it’s technically been a little more than a year if we were to include the time of serious development. A decent chunk of Eevee’s usability is also dependent on making 2.8 in general usable, and 2 years is actually considered a pretty short time frame when looking at production software as a whole.

In addition, I don’t see how Ton would not be able to build a better game engine when he oversaw the development of what would later become the BGE, the BGE is by no means a pinnacle of engine design that is hard to impossible to surpass. This being especially when looking at just how many Blender features it never supported and the large number of visual tools it still lacks compared to most of the alternatives (it didn’t even do WYSIWYG fully when looking at what was needed for effects like reflections). That is also not to mention that what you saw in the viewport did not always mean it was the same in the game (try getting an object to work with more than one vertex color channel as an example).

1 Like
123

https://www.nintendo.co.jp/support/oss/

I wonder if the switch will allow gpl games?

also, I can’t wait to get ahold of a mixed reality headset and try and get something going.

124

True, but it really didn’t fix some of the fundemental problems with the underlying funcionality of the engine…the many issues with the bricks and all the little things you cannot even change in official BGE like many of the wolrd settings you cannot change at runtime…we got some features that were needed yes…

I may have over stated my opinion, but what we did get was some icing, while the cake was still dry and moldy…maybe that is a better analogy.