Ton and other developers mentioned in the past that implementing a global nodal system will take several people working together about a year of work (perhaps the next code quest?) and Benoit’s work is a big part of the preparation of that.
That was for the “everything nodes”-project. This has a much more limited scope.
I think we are forgetting how the Blender Foundation, and it’s developers (and even the community), work.
Look at what happened with Brechts render engine. It was definitely not complete when he released it, in fact, it was more like a passion project at first, and it still technically isn’t complete, but he laid the groundwork for others to contribute upon. How many full-time, part-time, freelance, and volunteer developers now contribute and have contributed to cycles to make it what it is. I dont see this being any different.
Benoit most likely already has a pretty good idea of the design model he wants to develop against, he knows the Blender environment, he’s done this before, so I would say that it is likely he will hit the ground running. I would also assume that based on the schedule, he has other obligations, but the budget will support him to take some time to throw against this. As mentioned earlier, considering the previous engine was completely a passion project, I don’t think it is all that far fetched to assume he might not just clock out after 1.5 days on the dot.
I also agree with others in regards to how the existence of Eevee should expedite the process.
TL;DR:
I think this is exciting, as are all of the other super interesting developments that have been going on as of late. I think our time would be better spent supporting this project, and helping out when we can.
@Joel_nl well let us know, are you a developer too that you visit the institure or are you just living nearby in Amsterdam ?
I’m not a BGE user, and was kind of happy to see the old engine get removed, along with the Blender render engine because I knew they were both things the devs were unhappy to keep supporting and that were hampering the radical changes that are coming with 2.8. However, I think having some basic tools for interactivity will be useful in various fields (archviz for example) and, perhaps more importantly, will push devs to optimise Eevee so that it can be used with interaction mode, so I’m happy to hear it’s being worked on. However, I hope that it stays as simple interaction mode with full logic control over node graphs and doesn’t try to be a game engine specifically.
I’m curious to learn how far they want to push the new interaction mode, whether it’s going to be part of the nodification of everything post 2.8, and to what extent they will allow integration of other engines, such as Armory. I notice that Armory is already running as a 2.8 engine, so I’m really looking forward to hearing more in the next Armory update: https://twitter.com/luboslenco/status/1000055540280684544
and in 8 years this one will die of not funding the engine and code rot like the last one they tried to kill.
no thanks, I will be working on finishing upbge.
As you know, the design of the game engine was not optimal, because it basically duplicated lots of Blender functionality by having its own code just for that. The idea is to bring it closer, such that it keeps being maintained. I am also pretty sure that we are going to see the interactive mode being used in other contexts like previsualization, which would further help it to get some developer attention outside of games.
@Dantus, yes its maybe an underexposed thing of game engines, they’re great for complex animations too.
Complex animations sometimes can be done simpler with a game engine, i once did a transportation demo for a customer, where airport suitcases had to follow a trajectory of operations. the whole traject of sorting, switching lanes etc, spawning suitcases, collisions etc then BGE was ideal for that (key framing it all would be to complex). Game engines can easily act upon when something is near. And for the animator, it can become more like guiding a natural looking process, instead of trying to make animations look natural.
well I never made games, but i did several processing emulations, so its great 2.8 get it integrated as well.
also eevee would make it more usefull to me as well, for preview renders.
Wow completely missed this news https://www.youtube.com/watch?v=SP7qhKJZpds
Although not a independant engine, i think it seams allready how people would like game engines to be, node based or code based, edit in blender, play in blender compile play on dekstop or on web.
I talked to sybren (one of the blender developers in amsterdam) about it for a bit.
What was clear so far is that the new interactive mode (game engine) will be a fast/realtime engine (like eevee and workbench are engines) that will have things like physics simulation and baking to curves. But one of the main things is the ability for external game engines to directly link to this engine. In terms of features and input (controllers/touch).
So what it comes down to… the blender institure will not be building a new game engine from scratch, it will re-use allot of the work done from eevee, improve the physics parts and will create easy and stable ways for external game engines to link to blender.
So i guess we can all stop speculating now (for the most part) 
They learned a lesson from the old game engine, and thus using the graphics part from eevee makes sure they done have to implement the same feature multiple times as was the case in the past.
Ps: @BluePrintRandom thank you for your positive addition to this tread 
@Geographic The Armory3d developers did a great job.
With the work benoit will be doing other game engines can get the same (and better) results in terms of blender integration without having to re-invent the integration wheel 
I am glad Blender wont make or integrate any of game engines, instead they will make bridges to lot of famous engines. Blender 2.8 will be best 3D creation suite ever without bloat and code mess.
Times are changed, no point of making new game engine if there are already lot of good ones from Godot, Upbge to Cryengine and Unreal etc…
Any bridge to UPBGE may not work the same as with the other engines, as that engine is basically Blender 2.79 with a bunch of upgrades to the game creation side.
It would depend on how much the UPBGE guys will cut out of their fork so as to place all focus on game development (because otherwise, they run the risk of seeing code-rot in everything outside of the BGE code, save for a decision to rebase to 2.8).
As far as i am aware, upbge wont be cutting out any functionality. In fact we can backport all the good bits from 2.8, no need to worry about code rot.
The interactive mode is going to have additional advantages which won’t be useful for UPBGE. Dependency graph improvements are the main one. If the viewport or Eevee gets a performance boost due to a dependency graph improvement, the interactive mode is going to get it as well. That sort of improvement won’t help the UPBGE, because it is not using the dependency graph, but it’s own solution.
At around 2 minutes the Top Bar has a Game Logic tab showing.
That was before the BGE got removed from 2.8.
This is just the saved old layout… The name probaly just got stucked there… 
Ah yes, could be. Whoops.
What’s up on some .rpm installs for UPBGE?
Branch - eevee merge8
'it would be impossible to move upbge into eevee"
take that ton, right in your ton hole.