If it actually worked that way that’d be nice, but for example with Unity there’s some bugs that should be fixed in LTS but aren’t, so you’re still in the situation of updating to another version to fix a particular bug. It’s a huge logistical issue to try and maintain multiple builds and fix bugs between them.
The only way to solve all these issues is having the game engine developers re-design/re-write most of it to make it fully modular so that it could be possible to mantain compatibility between versions and having to update only some APIs and not the whole engine with all the APIs. For that to happen it means that a game engine software house must put a lot of effort and money obviously to make that happen in a short time (12-24months).
If Epic Games was developing a new UE5 engine maybe to be released in the next 24-48months then I would hope that a major re-design had been put in place long time ago to make it fully modular.
Anyway for marketplace products as well as any plugin or project a major design task to solve compatibility issues would still involve engine developers to create abstraction layers with auto-conversion that would just work with no bugs or issues. That would be possible but indeed very expensive time wise and so money wise because they would need a minimum of 5-10 developers paid to create conversion matrices tracking all updated and deprecated engine APIs classes/methods/macros and so on. It would be a huge task but surely doable in maybe 1 year or less depending on how many developers were assigned to it and how good they would be to achieve that.
Abstraction layers is the way to go along with full modularity in the long run with a major engine re-design.
Will that ever happen? Only Epic Games managers and CEO know that for Unreal Engine.
I must admit that made me laugh I’m aware of such articles I tend to keep up to date on what’s going on, but it doesn’t mean I’m going to give up.
What is the problem with upgrading your project to new version ?
we have a fairly big project, with (10-15 c++) plugging (most of them are custom ones/ internally developed ) and migrating all of this take max 3h from a version to an other ( so 3h every quarter is that bad… )
We wait until the 2 next version have received their last/third hotfix ( so we can leapfrop in case of a particular problem affect a version, without reverting )
exemple: we are in 4.20 right now, 4.21 has been fully patched, when 4.22.3 is out ( usually around the first release of 4.23), we will move to 4.21.3 and if any problem directly to 4.22.3
and by the way, UE4 is already quite a lot modular and they are still working hard to limite cross modules dependency
You can end up with cases where something significant breaks when moving to a new version, maybe something doesn’t exist anymore or it might break some work that you’ve already done.
Yes, exactly. If it were only 3 hours as Firefly says it obviously wouldn’t be a big issue, but there’s been enough cases in the past of things being broken beyond repair or take way to long to fix when trying to change to a different version, not even just code but materials and other things that it all adds up. The next big release is changing something with blueprints, so I wonder how many broken projects that may cause when upgrading.
Mostly they’re upgrading diff tools. Lot of dev-framework changelists is about that.
I don’t consider it risky. Especially that if extremely rare that new versions introduce big issues with blueprints.
Of course, LTS would be awesome. Although in the world of finite resources, I’d vote for continuing current philosophy of “rapid development”.
My opinion is, if a game is under development for 2-3 years you need a stable engine only in months preceding release. And even serious issues are possible to overcome thanks to source code access and a huge community.
The bigger issue for me is practice of abandoning “early access” features like material layers. You have no idea when they would resume development of such system since roadmap doesn’t exist.
Unity needs LTS so much because you’re unable to fix engine issues for yourself. I like this difference in development’s philosophy of the most popular engines, it gives us a choice.
Right now the engine is a mess given how fractured the bugfixes are through version 4.11 - 4.22 … the last 10 version of UE have all been extremely unstable (extremely compared to 4.8 to 4.10). I wouldn’t blame anyone who just sticks with an old version like me.
Currently I’d rather backport every 2 weeks bugfixes that affect my custom ue4.12 and have a nice smooth running engine. Rather than experiencing the ‘update your project to the latest version’ drama every 2 months where plugins and assets constantly break. And let’s not even mention the audience that’s not familiar with C++ or the engine itself… The way it is currently handled it hurts productivity.
I’d prefer they rather hire 2-3 more people for the next 3 years who maintain a LTS branch than giving away assets worth millions…
That’s not what that means, having to keep track of bug fixes between multiple versions of the engine is a huge challenge vs. doing new releases and hotfixes as needed.
Realistically you shouldn’t upgrade to a new version of the engine unless there’s a feature you absolutely need. Though it’s also easier to upgrade one version up than it is to upgrade several versions.
I wouldn’t personally prefer backporting because I like having new features.
That would be a dream come true if they make the LTS, and yes there’s a lot of regulars on the forums who often seem to forget that there is quite a few people not that familiar with C++ or the engine itself or don’t have entire teams of people dedicated to have a slightly easier time migrating their projects like others may be able to.
This makes sense on the surface, but when you take into account bug fixes, marketplace releases, tutorials, then it becomes a way different story and it’s the reason I decided I had to make this topic. It’s laughable excuse to me that it would fracture the engine to much to have a long term support version, the only real issue is that
[FONT=trebuchet ms]Epic Games needs to hire more staff members!
^ Last edited by Pine722; 05-17-2019, 08:11 PM. Date here for reference purposes I have a strong feeling I’ll need to be referring to having said this in future discussions over the coming years. More or less I’m “calling” it if you will. You know, counting the chickens after they’ve long been hatched, definitely not before.
I know you didn’t mention this darthviper so the following is not directed at you but I’m just going to make the following a general statement.
A lot of us are not Triple A studios who have been professionally trained/educated so it’s very tiresome dealing with some of the expectations that most of us are Triple A professionals who can easily handle constant engine upgrades etc and if not then tough luck sort of deal. Unreal Engine went with a free open source model and has a marketplace now, but the support didn’t catch up in all this time, I assume the support is the same it has always been for Triple AAAs who are paying for it but not enough people have looked beyond that to see, oh we have all these new non Triple A customers/users or solo indie devs who aren’t getting the support they need to utilize our tools. There may well be one indie dev out there who is satisfied in this regard but that doesn’t mean it’s the general consensus by far. Even if Triple A naysayers end up finding this topic I’m pretty confident that there are enough people out there who really want and would benefit from a long term support version if they had the option for it.
Also the fact that Unity has longterm support doesn’t mean it’s a better option for indies because it lacks a lot of other things such as quality marketplace content, a proficient visual scripting system, and more. Unity is lagging behind Unreal Engine in many ways otherwise it’d obviously be a better choice for indies & non Triple A because of its long term support.
As an example I’ve helped guide a few asset creators on certain topics & have encountered assumptions from some about things like ‘oh your modelers/level designers etc can help take care of this or that’, and I’m like ‘what modelers? What level designers? You do realize a lot of your customers don’t have many team members and may be doing everything or most things themselves.’ Basically the triple A only mentality. It’s still going on in the marketplace as well, you have some creators who expect you to have teams of people to fix or assist in setting up products that should have been ready to go from the start, like UVs, collisions, demo scenes, folder hierarchies, bug fixes etc. There are some prominent youtubers in the game dev scene that have openly discussed how Unreal Engine is currently not ideal when it comes to long term support.
Not to play the same tired card that has been played before, but if we take some actual facts that Epic Games is now worth "over 1 billion dollars", and has "the most successful game ever made", one MIGHT expect that it was enough money to hire more staff members to take care of certain things, and according to journalist articles, if they are at all accurate, it seems this is sorely needed because some people may actually be overworked at Epic Games, and I’m honestly not surprised that this could be the case, it doesn’t take a rocket scientist to figure out that Epic Games is severely understaffed for the amount of work they are trying to take on.
- It’s almost a year and there still isn’t a current Wiki since it was taken down.
- There also still is a severe lack of proper learning materials covering a wide range of topics that people need to make games.
- I still have multiple questions that I’ve asked on the forums for help with that have gone unanswered or unsolved, you know there’s a problem when not even the community is able to help each other & I do fully believe it is because of the lack of proficient learning materials.
- For example the resources for learning blueprints was posted in a sticky in 2014 and edited once in 2016, that seems awfully out of date to me, especially for an engine that changes versions 3-4 times a year. I doubt most Triple A professionals who are the few who have enough knowledge to help anyone, are going to spend their time on trying to help us on random forums topics.
- There is still no official Linux version of the launcher.
- We still are not allowed to change email addresses because of Fortnite.
- It took over 5-6 months, yes 5-6 MONTHS for the staff to fix a bug effecting me being able to use the marketplace on the website itself, and that is after submitting multiple forms, and then even some months went by during that time in which I was not able to use the launcher’s marketplace either, I was not even allowed to download the free content during this time because it uses the marketplace system.
11-28-2018, 11:34 AM
12-03-2018, 07:05 PM
I’m **still **waiting…
01-12-2019, 12:09 AM
05-17-2019, 07:08 PM
You can see both of my marketplace bug topics here:
12-03-2018, 06:51 PM
03-02-2019, 02:14 PM
04-04-2019, 08:14 PM
It still wasn’t fixed even some time after this, but I had stopped posting about it.
More examples of being understaffed involving other members of the community:
05-08-2019, 01:54 PM
05-11-2019, 11:31 PM
05-13-2019, 12:24 AM
**Whoever it was that finally managed to fix my account bless your soul and I really mean that, considering I know you are probably overworked to death. **We still wait for email changes though & many other things it seems.
Actually engine updates require much more effort from big studios (also so-called AAA) than small teams. It’s because of modifying the engine and producing a lot of systems/assets.
I’d say the effort it’s proportional to the project size.
@Pine722 None of the topics you mentioned is related to LTS… Although I’d probably agree with most of your arguments.
If you’re mentioning so-called “AAA mentality” here, I’m guessing you never worked on AAA in-house engine or even Unreal Engine 3. This generation of the engine is extremely indie-friendly. A single person can start making a game using this engine.
Hoverer creating games is complicated and this engine is designed to give freedom in building complex systems. Games need content, designers, and artists creating that content, so… I don’t understand part of arguments about “non-ideal engine”.
BTW, every gamedev company would use more people. Or every software company. Or every showbusiness company.
Not long ago indies we were supposed to download a bunch of unrelated SDKs and libraries and kits and etc… put them together somehow and call that thing a “game engine” xD
Everyone knew Unreal would cost a fortune if you’d want to buy a license… When Unity came along it was a big change.
Unfortunately Unity’s subscription model makes things complicated, I’d rather buy a permanent license like they had before (thus I left Unity).
Hobbyists can now expend a decade of personal free time on making a cool game with Unreal, if wanted. No pressure until you ship it, no subscription draining resources (per seat).
Unreal now “free to develop with” is crazy stuff, indies would never achieve the graphics of today without these open engines.
If it was bug-free, than that would be haven. But that isn’t possible to achieve…
I’d rather tolerate all those “services” issues than investing years putting a custom engine together, specially provided the fact I know zero about graphics programming lol
While I agree that it would probably take more effort for more complex games, the point being, as I previously stated ", “don’t have entire teams of people dedicated to have a slightly easier time migrating their projects like others may be able to.” multiple trained professionals having been hired to do those updates makes things vastly easier than for small teams/indie devs with no training.
ALL of the topics I mentioned are related to LTS. So I have no idea what you are talking about. There is a lack of “enough” support across all of Epics Services because they are understaffed, which brings me to the second point, because they are understaffed, they are not able to entertain the idea of a long term support version of the engine, hence things becoming to fractured for them.
I agree with you that this version of the engine is likely the most indie friendly compared to previous versions, but this is not the point of discussion I’ve been trying to make in this topic. The points I’m making are in regards to having a long term support version.
For your last point, If you are trying to say that you need to hire more people if you are a solo indie dev to work on all of these things, that completely goes against the idea of UE4 being indie friendly. Not everyone is in a position to be able to afford doing this.
Epic Games is constantly giving away free money for projects, free marketplace content, free games and signing deals but can’t hire enough people to give proper support of their products and services, this seems awfully odd to me.
"In December 2018, we announced Epic Online Services, a suite of cross-platform game services that has been tested in the heat of battle by 250 million players on seven platforms.
Since the release of the initial set of services in March, we have continued to develop its foundation. Our current efforts are focused on the immediate needs of store partner products that are approaching launch. We’re also working to open up the identity and social services we built for Fortnite to support cross-play in partner games without requiring the use of Epic accounts or other Epic dependencies.
As a result of our current prioritizations, **several roadmap items are being delayed. **Most notably, we are rescheduling the release of the Player Reports and Player Data Storage services."
Hopefully my points can speak for themselves.
Oh but they do try.
It’s not like Epic trying to save money, they have several open positions. They just can’t find “more bodies” to join…
I would argue programmers in game industry are overworked and underpaid, people look around and go work as a programmer somewhere else sooner or later, all benefits guaranteed and considerable higher pay, no crunch! no mass layoffs at the end of a project!
Just throwing more money at it won’t necessarily solve this kind of problem. Hiring fresh graduates won’t always help either…
They could say “okay let’s assemble a team to maintain this branch of the engine, only bug fixes + docs, no new features for a year.”;
Well first they have to find the people capable of working on and deliver that, may be the hardest part lol
I did forget to mention this in my post, so thank you for mentioning it, I know that the staff that Epic Games DOES actually have, is working very hard, and none of us would be here using Unreal Engine without their hard work. If it is truly the case they can’t find enough people to work for them, I think it would actually benefit them to discuss this openly with the community, it would actually benefit them to be more transparent about this because the community may be able to help reach out and find new hires. Our success depends on Epic’s success and in the long term Epic’s success falls on our success(although in the short term Epic may not realize that since Fortnite has helped them a lot).
Right now I feel like Epic is not communicating this well enough to us and so some of us feel shafted with lack of support, due to not enough staff. The transparency would help also because I don’t want to go by journalist reports only for this information, but in this case it’s all I have to go by, so I will say this, if it’s true that most of the people being hired are contractors and not safer permanent positions then this is likely the exact reason why they can’t find enough new hires. Again if this is true why do all the new hires have to be contractors? Even though they have enough money they are not willing to spend it on hiring non contract workers? Clearly this would be a big reason if it’s true, as people want to know they have job security.
Yes, communication sucks if you’re a solo or small indie who have no direct contact with Epic (UDN, conferences, evangelists, meetings) or you don’t interact with other experienced teams. Being based in Warsaw, quite successful gamedev city, makes these things a lot easier.
That’s what people are discussing a lot here, inefficient communication between Epic and “public”.
However, I’d strongly disagree with one of your statements. Let’s talk about it.
Nobody has been hired or trained for upgrading project to the new engine version. If there many programmers working on the project and somebody changed the engine’s code in a significant way, he or she performs merging. Smaller changes can be handled by whatever person dealing with general project management.
Blueprints usually require minimal effort since you can’t modify engine in blueprints
Issues with assets can be tricky, but in the case of UE4 the most problematic are Early Access systems like Niagara. That’s the reason why Epic is very slow with marking a new system as mature.
The same rules applies to upgrading engine as to any other development task. You stumble upon issue, you solve it.
If you know how the engine and tools work it makes your life easier every day. This is probably “the training” you’re thinking of, so I kind of agree with you here.
If you don’t know engine inside out and you don’t have the opportunity to learn from experienced devs on a daily basis, I’d push for better communication and docs. For instance, there were few great talks about blueprints and asset optimization during Unreal Fest Prague a month ago. Still, they didn’t post any footage…
I don’t think it is an issue. The game Sea of Thieves was made in UE4.10 and we were already at 4.18 at that time and this didn’t prevent them to have a game launched. They did made from source and added fixes released later during development, so it was their specific branch lets say.
If someone would start doing a game today (day zero) they could start by checking if there is any UE4 release that would have everything they need in first place, then check if the most important aspects which matches their game style and mechanics are compromised by any bugs. If there are bugs, they would check which is the planned release of those fixes. Also, it would be interesting checking new techs (ie Niagara) and when that would become matured and you can expect that will become a solid tool, since it is in tests for long, it will just take more time because the recent realtime ray tracing included recently otherwise it would be released already.
The engine, as it is, is only part of the equation on producing a game. There is so much more work besides that: models, texturing, environments, and all of these are not affected by any engine bug considered a showstopper.