Ditch unity and rewrite at unreal engine

[Jeq]

This is my dream and it has been dream before early access game. When EA came i were even more sad it had to be at unity.

It can be done fairly fast. Codemasters moved from their own game engine which dirt rally and dirt rally 2.0 presented and now they developed WRC 23 at unreal engine.

Beneficts for this would be better graphical performance and optimizing because of nanite. 

And i think this way you would get away from some physics bugs which Unity have, which cannot be fixed. This is only way to slay kraken.

Edited September 18, 2023 by Jeq

[Periple]

I wish it was that simple but it really isn’t, or we’d port all of our games to Unreal too!

 

[The Aziz]

1 hour ago, Jeq said:

It can be done fairly fast.

Wrong.

1 hour ago, Jeq said:

Codemasters moved from their own game engine which dirt rally and dirt rally 2.0 presented and now they developed WRC 23 at unreal engine.

Dirt Rally 2 is 4 years old.

1 hour ago, Jeq said:

Beneficts for this would be better graphical performance and optimizing because of nanite. 

And that's about it. Do you know how well Unreal handles complex calculations required for the type of game KSP is? Of course you don't. Meanwhile completely omitting the mountain of work required to switch engines, get a new team experienced in such engine... Pretty much rewriting the game from scratch. Do you really want KSP2 EA 0.1.0 in 2035?

1 hour ago, Jeq said:

And i think this way you would get away from some physics bugs which Unity have, which cannot be fixed.

Yes, and it would introduce physics bugs Unreal has.

Stop the fantasies, it ain't gonna happen, not this far into the development.

[Spicat]

• This will probably solve nothing as the major part of ksp problems are not delivered by the engine.
• Performance problems are mostly CPU based and Nanite won't solve most of that (Better solution if you want to start the game from scratch is base the game on the new DOTS tools of Unity)
• Developers know Unity better so you would need to hire new ones that are more knowledgeable with Unreal.
• Unreal is much less moddable than Unity, so goodbye most of your mods.
• You will need to start the game from scratch, I don't think the community will be happy with a 6 years delay.

Edited September 18, 2023 by Spicat

[Superfluous J]

I'm pretty sure you can just Export from Unity and Import into Unreal. Make sure you check all the "high quality" checkboxes first, of course.

 

 

 

Spoiler

Just in case you don't know this from reading it, I'm joking.

 

Edited September 18, 2023 by Superfluous J

[InterstellarDrifter]

1 hour ago, Superfluous J said:

I'm pretty sure you can just Export from Unity and Import into Unreal. Make sure you check all the "high quality" checkboxes first, of course.

 

 

 

  Reveal hidden contents

Just in case you don't know this from reading it, I'm joking.

 

I was about to say....

[KerbalsRus]

Maybe just get Chatgpt  to rewrite it 

[HebaruSan]

15 hours ago, Jeq said:

It can be done fairly fast. Codemasters moved from their own game engine which dirt rally and dirt rally 2.0 presented and now they developed WRC 23 at unreal engine.

Your example to prove it can be done "fairly fast" is... a game that isn't released yet, promised for two months from now? Which, depending which sources you believe, may originally have been planned to release in July? And for which you do not state when development began for comparison, perhaps because no such date is publicly known? Am I getting that right?

Edited September 19, 2023 by HebaruSan

[Gotmachine]

Aside from the fact that switching from Unity to Unreal would basically mean restarting from scratch and require several years to just come back to where they are now, the whole idea that the game would be better by doing so is questionable at best. The various cutting edge stuff from Unreal (you mentionned nanite) is more or less useless for a game like KSP. The only significant dependency KSP has on the engine is rigidbody physics, and while it can be argued that the Chaos engine from Unreal might be a slightly better fit that PhysX in Unity, the best fit would actually be either a custom implementation, or at the very least a forked version of an off-the-shelf physics engine.

Most technical (and performance) issues in KSP 1 or 2 comes from subpar technical solutions due to a lack of an adequate software engineering capacity given the scope and design requirements. The engine they are using is a minor factor.

If they end up miraculously getting greenlighted for a large refactoring project, it's more likely to be a project revamping their core architecture issues in the light of the experience and feedback they acquired in the last years than a base engine switch.

[Dakitess]

I'll be sincere, I'd love an official statement saying that they'll start again from scratch on a whole new modern engine, even if they announce it to be EA in 2030.

I don't see KSP2 being worth the "2" before 2 good years at least, and it would still be very very limited compared to what a "KSP2" can / should / must be. And it very well might be a final point to the KSP History which would be unfortunate.

So yeah, since I can't expect any serious KSP3 being developed, even for 2035, because of the massive failure of KSP2, I would really be hyped by a "Sorry, ignore everything of the past months, we'll start it again, see you in a few years, you'll trust us again". It would at least give some hope about KSP being something for coming decade, rather than this drastically diminishing interest, still going down, and taking down KSP1 with it unfortunately.

This is not sardonic by any mean, just a true feeling, I would really see this as a good news.

Edited September 26, 2023 by Dakitess

[moonme]

Unreal may offer some benefits, such as more advanced solid state physics, but those benefits are likely to be minor. I think the best strategy  would be to focus on improving the existing KSP code.If they improved solid-state physics, it'd already be cool

[Periple]

2 hours ago, Dakitess said:

I'll be sincere, I'd love an official statement saying that they'll start again from scratch on a whole new modern engine, even if they announce it to be EA in 2030.

This is pretty squarely in "can't happen" territory. If KSP2 doesn't fly for whatever reason, they'll just close down Intercept and mothball the IP. I would really only expect that to happen if T2 decides to close down Private Division. In that situation, the best we can hope for is a spiritual successor from somebody else.

(Moreover, the engine really isn't the problem, and there isn't another engine on the market that's obviously better suited to make KSP2 than Unity.)

[Dakitess]

Yeah I don't see that happening either, for the reasons you mentioned... Ha, let me dream of a KSP3 then :p

[cocoscacao]

6 minutes ago, Dakitess said:

let me dream of a KSP3 then

I'm getting off topic, but... does anyone believe there will be KSP 3 (assuming KSP 2 turns to a resounding success)?

[gluckez]

2 minutes ago, cocoscacao said:

I'm getting off topic, but... does anyone believe there will be KSP 3 (assuming KSP 2 turns to a resounding success)?

ksp2 hasn't even released yet

[Dakitess]

Because of the state of KSP2 and the very bad momentum it gave to the franchise and the community, no, I don't see that coming. Which is what is heartening me the most and why I speaking about "dreaming" than hopping : KSP2 failure would not mean a bad KSP, but the end of it, or barely, while damaging the first one. Dang.

I don't know if getting a "proper" KSP2 by 2025 would be enough to ensure a KSP3, be it in 2035 or more. Like, if a roadmap-finished and somewhat working KSP2 but not fantastic by any means, would help turning the page. I feel it needs to overcome the status of being simply "finished" and completed, with all it went through, or no one would ever risk trying a KSP3 dev.

Edited September 26, 2023 by Dakitess

[Periple]

1 minute ago, cocoscacao said:

I'm getting off topic, but... does anyone believe there will be KSP 3 (assuming KSP 2 turns to a resounding success)?

Microsoft Flight Simulator has had lots of versions so why not KSP? If they can keep selling updated versions of the game, why wouldn't they?

[cocoscacao]

Just now, Dakitess said:

Because of the state of KSP2

3 minutes ago, cocoscacao said:

assuming KSP 2 turns to a resounding success

 

[Periple]

Just now, Dakitess said:

KSP2 failure would not mean a bad KSP, but the end of it

Yeah pretty much, the franchise won't work without it and it's not big enough to start over a second time.

[cocoscacao]

2 minutes ago, Periple said:

Microsoft Flight Simulator has had lots of versions so why not KSP?

Never really checked player numbers, but KSP is kinda obscure game

[Kerbart]

Practically none of the issues is caused by Unity. The few that are, like the registry bug, are caused by devs not reading the API documentation and that can happen with Unreal as well. In fact, given the unfamiliarity of the devs with Unreal, it's even more likely. Switching to Unreal wouldn't bring any benefits.

[Periple]

MS Flight Simulator has about 10k peak players. It's not THAT huge, and also it's a more contested genre. I think KSP has the potential to be bigger. Factorio f.ex has about double that.

Edited September 26, 2023 by Periple

[cocoscacao]

2 minutes ago, Kerbart said:

Switching to Unreal wouldn't bring any benefits.

So switching to Unreal, is unreal. I think there's a paradox hidden in here....

[Kerbart]

3 hours ago, cocoscacao said:

So switching to Unreal, is unreal. I think there's a paradox hidden in here....

Yeah but to say that the community is united about Unity is disputable as well...

[The Aziz]

17 hours ago, cocoscacao said:

think there's a paradox hidden in here....

Please don't come out of hiding Paradox, I don't want 38195716497 dlcs for one game.