OPINION: How to Make Lumberyard Better?

Sergey Korolev discusses some ways Amazon can make Lumberyard a little better.

Sergey Korolev talks about his search for the perfect engine and discusses some ways Amazon can make Lumberyard a little better. This is an opinion piece, so keep it in mind, when you’ll start killing people in the comments. 

Introduction

Greetings! I’m Sergey Korolev. For over 10 years I’ve been searching for a perfect game engine. That’s kind of a hobby. I’ve helped to promote Unity3D in my country after its launch. Provided an official support, sales and education to Virtools engine users before that. And of course I tried building my own engine long ago.

Today is a great time. Because we have Unity 5 and UE4. Great features, communities, marketplaces, tons of tutorials. You can do everything you want for free. And, thanks to that new price policies, many other engines have hard times now.

There is still place for other engines. They just need to change. Yep, they should be really powerful. But if you know what game developers are searching for – it’s not a problem. For example, in my company we are using 3 different engines (depends on genre, platform, etc). One of our developers worked on Warface project.

In the end of 2016 Crytek released CryEngine 5.3. That year was hard for company and they closed few offices. Also, they have sold CryEngine 3 to Amazon in 2015 (I believe that even Stonetrip Shiva would be a better choice). Have the things changed after that? Nope, the new free CryEngine chooses a strange and unstable way to go. So let me guess, 2017 will be hard for them too.

In January 2017, after 1 year of development, Amazon released Lumberyard 1.7. Beta for some reasons. I tried it again to look at changes. And it makes me really sad. Amazon has so much powers, but for some reason that project is going the wrong way too. They have a nice marketing and resources on one hand, and a really bad tech and support on the other. There is a big chance, that engine will be eliminated as a bad experiment of the company.

So let’s discuss if it’s possible to fix these engines. And talk about the reasons to use them.

Lumberyard quick start

I think it’s a good idea to start from scratch and show you what I mean, because usually you can see awesome pictures like this one:

As a new user, first of all, I will check game engine’s features list and its price. Just look at the site.. free, AAA, funny beaver logo, etc. Then I will download installer and try it. Lumberyard looks really confusing:

Yep, this is Setup Assistant. I’ve reported about problem with fonts in summer 2016. And it still hasn’t been fixed. By the way, it was ok in previous versions, so I can understand how it works. Idea of that tool is good. Just check what you need and it will be installed.

Let’s move forward. Start Lumberyard Editor aaand wait.. for 5 minutes in my case (with Alienware Alpha and SSD). That happens thanks to the annoying Asset Processor. And nothing like that exists in CryEngine 5. So I believe it’s possible to optimize it somehow, and that should be the high priority target.

Finally it started. Now it’s a good time to see some impressive examples. Usually I’m just trying to find some powerful template or something like that, just to see the possibilities. 

Wait. What? Ok, there is some ugly Android demo, and other examples like that. Of course you can try to use GameSDK (shooter) project, but it looks totally broken now and I don’t think that it will be fixed (even Crytek doesn’t plan to support it anymore). Hey, what about AAA? How is it possible?

That all looks like an engine of 2005. And what’s happened to the assets? It’s a total chaos right there (on the left). Tools usually look really old, it might be painful to use them. But let’s forgive that moments, because Amazon is honestly trying to replace them, and it’s a good thing.

As a new user I want to compile the game and get a standalone build. But you will not find a “Build” button here or in the CryEngine. And also project assets are not separated from game engine itself (there is a nightmare with files too). So if you wanna make a standalone project be ready to feel a real pain and follow the guide like that:

Sure, it’s better to read a quick start guide first. But usually users are trying to use it exactly as I’ve shown. Install/start/open examples/trying to create their own test project/make it standalone. And only then reading some Docs. 

As result – there is no community. But that is the core of any engine. Community can make assets, help you make or fix something (while developers usually answer too long), push technology forward etc. No community – no future. Because game engines today are a kind of free2play service, not only just technical base. Answers on forums aren’t enough. The engine should be friendly itself, integrated with a modern community features and offer some awesome templates.

Here is just an example of typical user test:

That first impression should be good and first steps should be as easy as possible. Yep, it’s just a basic overview and there is many other aspects. But I’ve honestly spent over a month trying to make something…and it wasn’t fun at all. Nothing works as it should, bad support, not enough docs etc. First week just to understand how to make a standalone version, and unlimited time to make a dedicated server work.

Sure, you can make some games with it. Spend time, learn it, read forums. But why should you do that? Why choose that engine, if there are better alternatives? Yes, it’s only beta. But Unity3D/UE looked better even in this stage. Because it’s just a question of developers’ philosophy and priorities.

In 2005 Unity3D was available only to Mac, and become really popular after Windows release (but it was great even on the early versions)

Comparison

To be honest, CryEngine 5 is not really better. Some core things are fixed, of course. But you can post a critical bug and it can be easily ignored. Each update is not stable, community is not really bigger, support and Docs are bad, etc. Ok, just take a look an that classical game engines comparison spreadsheet (I know you’ve been waiting for it):

Any questions? Any reasons to use Lumberyard or CryEngine today? Sure it’s just my opinion and that comparison might not be totally right. But the whole picture looks exactly like that for users today.

The first one is still far away from Cry Engine 5, trying to become kind of Unity (with Editor UI changes and Components system, for example). But CryEngine 5 looks like a broken version of UE3 today. So they are both behind modern engines now. Ok ok, in some cases you can use it.

In Lumberyard case – that is Amazon itself, AWS and Twitch integration. Awesome idea, but without strong technical base – it’s all useless. You will not find much games, made with Lumberyard, but there are some interesting titles in development now. Recently Star Citizen project migrated to Lumberyard. Well, I think the only reason here is some financial support from Amazon (looks like they ran out of deadlines and the budget). Because otherwise it’s insane and I’m not sure we will see the game release or even beta in the near future.

Star Citizen moves to Lumberyard (that will be a looong trip)

In CryEngine case – today the main reason is only GameSDK. Because it’s kind of modding and you can save time thanks to that. But it’s really unstable after each update (tone of bugs and buggy editor). Crytek does not plan to support it and wants to remove LUA support (that is used everywhere in GameSDK). And if you want to make a multiplayer shooter – think twice. Because the main advice that you will get from developers now is: “Do not use it, please”.

Just take a look at Miscreated, for example. Developers complain about game engine quality (please give them a medal for what they’ve already done). Especially about Networking (have plans to fully replace it). I think that is the reason for empty streets in this game, because it’s impossible to extend the world with current game engine optimization and networking. It’s just too old for modern games.

Miscreated was presented in 2016, developer are trying to smile and not to cry.

Also, just check release notes. You can wait for the better times and really important fixes (even not features!) really long. Because usually updates are rare (1 per 2-3 months), small, with strange highlighted features like that:

Everything is worth, but let’s make that Spawner component first!

Or just check the roadmaps. Lumberyard plans to increase Mobile (does it really need mobile now??) and VR support, make an FBX Important, add Twitch ChatPlay etc. Don’t wait for anything really serious from CryEngine too:

Here is a fresh release notes for you to compare:

How to change the future

I know, that my feedback looks critical. But I just hope that developers of both engines will change current direction or that I’m just wrong and they have a better plan.

Why bother? Because developers can make a mistake, choosing these engines today. Because I don’t wanna see only 2 leading game engines on the market (developers should have a better choice). And both technologies still have a good potential.

So let me offer some possible changes for both engines. Sorry, no pictures for you here. I just wanna make that list as clean as possible.

Here is the 5 steps to save the situation and change the feature:

1. Make it easy to install, test, build and deploy

Make an installation and first steps as simple as possible. Do not send users to Docs – integrate tutorial and your advices right inside your editor (if that is really required). It should be as easy as possible to create a basic arkanoid for example and then share it as a standalone version.

Most developers today are kind of casuals. Yes, you can focus on the serious companies with a big resources. But ignoring indies may be a big problem. Just take a look at the Harvest Moon.

Yes, they might not give any profit usually. But they are really active, and can help push things forward. Game jams and funding are perfect for that.

2. Focus on your community and their requirements

As I’ve already said – community is everything. Build and extend it. Try to listen to your users for real. What are they waiting for? Why should they stay and choose your engine?

Usually engine developers just talk and talk and follow their own way. But It’s not enough to just talk. Change the roadmap, depending on the user requests.

First of all, core features, stability and convenience. And only then additional services, mobile, VR, modern rendering or whatever.

3. Support as fast and as good as possible

The most horrible thing is ignoring tickets or issues from users. Or just answering too long. Especially if the problem goes from engine issues or poor documentation. As result developers will stuck and hate or leave game engine.

So just hire more people to support. Answer as soon as possible. Docs should be always fresh and overextended. Users will be thankful for that and continue to work with your engine, despite any problems.

And the main thing – extend your QA division. Test each step. Run your templates and examples again. Is everything fine? Are you sure? Do not leave a critical bugs to the hotfix. Because it’s hard to make a bug-free game with game engine full of it.

4. Become a modern service

Ok, so you are making a free game engine. But you still need a profit, right? So make your interests transparent to game developers. No secrets. They should understand why it’s free. Because for now it looks like it’s free, do whatever you want with outdated Docs and poor support.

Maybe you shouldn’t be Unity or UE4. Find your own audience, own path. Make a flexible license, take money from developers’ profit. But save quality, and speed up development. Your users are ready to pay for that. Or make good and attractive services like Unity3D does. With Twitch and scalable integration – that might be a great solution. Plus add more collaborative features.

Of course, the asset store is important too. But if there is no community – it’s useless. It is a good idea to hire a team and make some awesome starter content. Like Unity3D does with Adam demo for example.

5. Create the game for modding

And here is the main thing, that can make any engine super attractive in my opinion. Today the game engine itself is not enough anymore. Because we, game developers, already have it.

Engine developers must create something more than simple templates and then share it for free. The best situation would be something like Source Engine. You should know how many games were made, based on it (and that’s why I’m putting so much hopes on version 2). 

Now you can understand why I’m still voting for GameSDK (before another good alternative). You don’t need to create a game from a scratch. It’s more like a modding. Sure there are different game templates on Unity and UE marketplace. But usually it’s alright just for prototyping (thanks to the bad architecture, optimization or just very basic features).

And I believe that it’s the future of the game engines. Usually it will be like a modding with a different additional building blocks that are easy to integrate and use. Even AAA games. That future is closer than you can imagine.

Is it really possible?

Sure. It’s not an advertisement (Or is it? Doesn’t matter), but just take a look at Defold engine. Yes, Unity3D is kind of a standard for 2D and mobile today. But it doesn’t matter to Defold, because that guys are just trying to take their own segment.

I think that’s all possible thanks to Oleg Priduk (post-evangelist of Unity3D). When he came, King’s game engine became much more friendly and attractive. More tutorials, good community, game jams, good roadmap and everything that I’m talking about here. All in the short time period. 

Summary

I’m not recommending to use any of these engines today. But they still have a high potential and I hope that some of them will find the way to realize it. So let’s wait for a better version of Lumberyard and CryEngine 5.5.

In some cases, it’s possible to use it, but you should be careful with your decisions. Sure, you can see the profit in that choice, but I’m not sure that your team, investors or boss will be thankful in the end. But if that works for you I will be happy to hear your “hey dude, you wasn’t right!”.

Both Lumberyard and Crytek team are trying to make great things. Very positive guys, interesting ideas. But they are really at risk right now and need to change something. Anyway, I wish them a great luck on that hard way.

Sergey Korolev, Game Developer

Join discussion

Comments 7

  • TinFoilHat

    Lumberyard discontinued support for linux and Vulken choose mac/ios with metal. Now Valve has Vulken running over metal and it's MoltenVK. Amazon has made the same mistake that godot has made.Godot started working on Vulken the next day after they saw Moltenvk working . I'll never use any of Amazon's services. Mobile first...

    0

    TinFoilHat

    ·6 years ago·
  • Dingus

    I have never seen a more idiotic "comparison" table with my eyes. Really makes you question the competency of the person who made it.

    I do agree that he has really valid points about Crytek not focusing on its dev community. This really needs to be rectified in order for them to stay afloat.

    0

    Dingus

    ·6 years ago·
  • Binky

    Hey Sergey, we've talked but I wanted to say thanks for writing this well thought out and insightful article.  You make a lot of valid points and the team is taking them to heart.  The customer and community are very important in our decision making process and it's this type of feedback that helps guide our ship.  I'd like to point out our community is growing (we are still very young) and fairly active.  You can check out the forums here:  https://gamedev.amazon.com/forums/index.html  (Note:  we are in the process of making some major changes here).  

    Please keep the feedback coming!

    Chris "Binky" Launius - Lumberyard Community Manager

    0

    Binky

    ·7 years ago·
  • sunnlok

    We also have a very active community going on in the form of a slack chat where crytek emplyoees and experience community members try to help and talk with everybody who comes over. It's quite active and you will certainly find someone to help you with your problems.
    Chat: crycommunity.slack.com
    Reigstration: http://5.45.111.147/chatsignup/

    0

    sunnlok

    ·7 years ago·
  • Patrik

    Iam just a hobbyst but iam following CE and its developement since Crysis mod sdk. Crytek had some bad decisions in the past, but i see that they are really heading to the right direction now. Maybe they need some more time to catch up with UE and Unity in some areas but they are defenately getting better every day. And Collin really did a great job on video tutorials and helping others. You can also often get answers on the forums from him and other Crytek guys. They are really helpful.

    0

    Patrik

    ·7 years ago·
  • Anonymous user

    Thanks for the details, we'll update the spreadsheet.

    0

    Anonymous user

    ·7 years ago·
  • CollinBishop

    Sergey,

    Thanks for the piece on CE/Lumberyard. I can say that we at Crytek have been overhauling quite a bit. We are really taking community issues to heart and tackling many things at once.

    A quick note on the spreadsheet you posted had a few things missing in it for CE. We have had a complete overhaul on the Particle Editor and it is now node-based. We have also gotten FBX importing for static and animated assets in the engine as of the 5.2 release. And to just bring awareness to the video training we have added around 200 video tutorials on the new editor/asset pipeline to bring new users up to speed.

    Honestly though thank you for the breakdown and know that I will look over the points you made over the coming weeks.

    - Collin Bishop (CRYENGINE PM)

    0

    CollinBishop

    ·7 years ago·

You might also like

We need your consent

We use cookies on this website to make your browsing experience better. By using the site you agree to our use of cookies.Learn more