Download

UE5 for active development

Epic said that UE5 is in Early Access and not to be used for active development of projects.

Is it me or does everyone seem to be using it for their projects anyway or are they just testing it?

1 Like

I havent got anyone else’s money tied up in my UE5 project, no specific deadlines. If I hit a blocking issue its not a big deal to wait for a fix, and keep working on other content and areas of the project. I’m not likely to be releasing anything before Epic get done with version 1.0 of UE5 so I’m not too worried about that.

All that aside, 5 seems to be built on the fairly solid foundations of 4. UI has changed a bit and there are some incredibly promising new features under the hood, but it feels familiar. That mixed with the lack of a deadline and how visually rewarding some of these new features are, means its hard to pass up just going feet first with a project in 5.

If I was working a 6 - 12 month turnaround with other guys’ money on the line, the known quantity and stability of 4 might be a better option. For beginners, there is a lot more documentation on features in 4, and tutorial content will be more reliable and easy to follow.

With time and experience on your side though, no real reason not to have fun and play with all this new stuff. Lumen is worth the price of admission alone when you get to grips with it a bit, even with its teething problems.

I imagine most here are of a similar viewpoint on it.

1 Like

Yes, I tried UE5 myself and it was great how familiar it was straight away.

The problem for me is that those tempting features can cause you to start active development in UE5 but if you have spent tonnes of time on the project in 5 and some upgrade happens later that breaks projects then it is going to be wasted time, especially given that Epic said to not use it for active development.

It’s safer to stay with 4 with a roadmap that will eventually lead to 5.

UE5 is causing a lot of dilemas and uncertainty given the lack of clear information and direction with regards to landscapes, virtual texture/heightfields etc.

The risk you mention is certainly worth considering, and I don’t think there’s necessarily a ‘wrong’ choice - its going to be dependent on a lot of specifics about a given project.

Epic state not to use it for active development thats true, but they also state that what you do in Early Access will/should translate to the final release as much as a 4.27 project should. I can’t see that I’m necessarily taking any bigger of a risk per se; in either case I am relying on them making that transition work as well as possible on their end.

If your project is relying heavily on landscape/virtual heightfield stuff I can see why you’d stick to 4 for now though. A lot of the good stuff in 5 is not really working in those areas at the moment and as you say, who knows how much will translate in the end, or what form it will take. There’s definitely the potential for things to break and for time to be lost - thats not necessarily the case in UE4 where if you dont like the prospect of what would be required to move to 5, you simply finish it in 4 where the feature set is stable and unlikely to undergo major changes.

My project is agnostic enough to those issues that I can feel comfortable going ahead in 5, and if I have to rework some stuff down the line on account of it, then I guess I’ve just mentally accepted that risk. Naturally I hope that Epic are simply offering a disclaimer in case major changes are necessary, so that it can’t be said that they misled anyone, and that the transition to full release will be relatively painless - but this is a best case scenario and I am prepared for the possibility of not getting what I want for Christmas.

As I said above though, lack of deadlines and a personal appetite for risk and the bleeding edge of visuals make that something I’m personally OK with. All I can do is try to mitigate the risk as much as possible by focusing on developing aspects that I consider ‘safer’ on the sliding scale of features. Landscapes/VHM would definitely not be one of those things and if my project relied heavily on them, I’d be more wary for sure.

Good thread though, food for thought for people mulling over what direction to go in with this.

2 Likes

If I wasn’t using it on an unreleased personal project that highly benefits from the new features I wouldn’t be using UE5, never mind the ue5-main branch. Just like you wouldn’t want to be using preview builds or the master branch.

It can work a lot of the time. It can also break in unusual ways, very badly. UE5EA is largely 4.26 and in comparison to ue5-main, it’s even more outdated now. Right now It’s just going to cause headaches or redoing work because you can’t go backwards in versions.

If you want to just actually work on a project, you are best off sticking to the UE4 releases and upgrading later.

I wouldn’t. I say that, but i did for a while and it wasn’t a fun experience. Crashes left and right.
Basic features (like foliage scalability if you do certain things) instantly crash the editor when enabled.

No idea what glossy brochure you’re reading. Devs on here are wondering if 4.27 can even be used for active development!!! :stuck_out_tongue: Seriously, we’ve still got show-stopping bugs like D3D Device Lost and yet no official reaction or comment from Epic on the official forums ever. How can that be? I don’t get it… So I decline any upgrade until there’s more transparency…

1 Like

You sacrifice stability and stable features for the luxury of dipping your toes with the next big thing. If you’re a big studio, I would assume you get a more stable version of UE5 but big budget games are being made with UE5 as we speak. They know what they’re in for and presumably work directly with Epic throughout the way which is something we don’t have. After all, Epic needs to showcase games with Nanite and Lumen :wink:

Personally I don’t take part of experiments or beta tests for anything until it has proven itself to be stable or safe to use. I may like the idea behind a self driving car but I’m not going to buy one and be a crash test dummy for Elon. Hopefully we’ll see a stable enough build by 2022 but so far UE5 is fun to play with.

Yes videos are being highlighted of projects using UE5 and that makes me as a dev think about using it because the bigger studios are. This is why I made the point about “everyone” seems to be using it. But this is at odds with the “not for active dev” notes.

It’s enough of a dipping of the toe using UE4. It has tonnes of features and is a hard beast to tame sometimes. UE5 is like diving straight in the deep end and if it’s a production project then there aren’t any ladders to get out with. :sweat: