Why OGRE is Dying

[Kojack]

When it comes to building the engine, we recently added quick start build scripts for 2.1 & 2.2 to handle that step in a few clicks.
...
I hope the quick start build scripts help you with that!

I tried them out today, they work pretty well. Although I ended up spending something like an hour stepping through the code because every sample failed because the d3d11 renderer couldn't be loaded due to the module not being found, despite the renderer being there.
Turns out it was because the build script didn't copy the amd_ags_x64.dll library into the bin directories. Damn windows error message just said module not found, it doesn't say which module, and ogre only knew the renderer wasn't loading.
I copied the lib and it all works now.

Otherwise, far easier than it was.

[xrgo]

I also just tested the build script, worked at first try, but also had to copy the amd_ags_x64.dll file by hand to run the samples

[hydexon]

Since was a loong time i've posted here, i want to post my thoughts.

Many people as well newcomers, they mistake OGRE as an fully-fledged game engine instead of an rendering engine, this is because many blog posts, forums across the internet treats them as game engines (Google is one of them when you type "game engines" and Wikipedia as well).

Also i've loved the progress of OGRE 2.x series, but the legacy codebase 1.x cripples them too much to being adopted as modern rendering engine, for example the extreme dependence to the ResourceManager which many find as obstacle, like me which i've wrote an asynchronous asset loading/processing system, but i had to discard it because the OGRE resource manager doesn't like multithreading.

Many of them is moving to other rendering engines which supports the newer APIs (DX12/Vulkan) and are much more lightweight and non-intrusive: bgfx, DiligentEngine, TheForge (this is an kind-of exception since their build system is quirky and messy, but they even support game consoles such XboxOne, PS4, and Nintendo Switch upon request).

Despite i've liked the OGRE's new improvements, my engine architecture was not compatible to OGRE,since used raw Vulkan to test multithreaded environments.

My two cents.

[longer]

OGRE not dying but survive difficult.
OGRE is only a rendering library. Just a toolset. This decision is successful but painfully.

Mobile platform is a big opportunity, But OGRE 2.x does not support well.

I think OGRE lacks a common component across platforms:
```
0. Vulkan (API) for mobile platform is Urgent.
1. Unified file system, like cocos2d. At IOS and Android have same interface.
2. OS System event, like touchs event, I think ois is not a good choice. We need native solution.
3. Embedding OGRE into Android ios need more easier, The launch process of OGRE is too vague,
Especially RanderTarget mapper to Android SurfaceView, IOS UIWindow UIView.
4. CMake is good, but A specific compilable project is so important. Many newbies are confused here.
In addition, when I use ogre, I manually re-do all project for the dependencies and ogre itself,
so my partners will not get lost in CMake.
5. Resource format export tool, mesh material etc. The blender2ogre is good, but not Official. not stable.
Resources are more difficult to produce.
6. UI Library, MyGUI CEGUI recent update seems to be slowing, which makes me very worried.
Once the downstream ecosystem of the rendering engine stops, it brings the loss of users.
```

Cross-platform out of the box is very attractive.

This is my build system in a specific environment, which can be compiled directly and completely.
https://bitbucket.org/mm_longcheng/mm_lib

[paroj]

OGRE not dying but survive difficult.
OGRE is only a rendering library. Just a toolset. This decision is successful but painfully.

Mobile platform is a big opportunity, But OGRE 2.x does not support well.

I think OGRE lacks a common component across platforms:
```
0. Vulkan (API) for mobile platform is Urgent.
1. Unified file system, like cocos2d. At IOS and Android have same interface.
2. OS System event, like touchs event, I think ois is not a good choice. We need native solution.
3. Embedding OGRE into Android ios need more easier, The launch process of OGRE is too vague,
Especially RanderTarget mapper to Android SurfaceView, IOS UIWindow UIView.
4. CMake is good, but A specific compilable project is so important. Many newbies are confused here.
In addition, when I use ogre, I manually re-do all project for the dependencies and ogre itself,
so my partners will not get lost in CMake.
5. Resource format export tool, mesh material etc. The blender2ogre is good, but not Official. not stable.
Resources are more difficult to produce.
6. UI Library, MyGUI CEGUI recent update seems to be slowing, which makes me very worried.
Once the downstream ecosystem of the rendering engine stops, it brings the loss of users.
```

Cross-platform out of the box is very attractive.

This is my build system in a specific environment, which can be compiled directly and completely.
https://bitbucket.org/mm_longcheng/mm_lib

0. How is that urgent?
2. OGRE uses SDL2 nowadays, as OIS is indeed not a good choice.
3&4. Did you see that we provide a prebuild Java SDK of Ogre for Android: https://www.ogre3d.org/download/sdk
5. blender2ogre is official: https://github.com/OGRECave/blender2ogre Also there is OgreAssimp
6. we recently incorporated Imgui into OGRE: https://www.patreon.com/posts/important-coming-30677960

[longer]

Sorry, my bad English.

I work for ogre for mobile platform need support windows android ios.The windows use GLES2 for development and debugging.

I use ogre 1.x because i need mobile platform,I can't upgrade to 2.x.

I see the sdk but not use it.I can not commit a sdk to a repository.When i use the cmake generate a project.Tt is look confusion, especially the windows project use absolute path.This brings inconvenience to my partner.When i fix this,i create a new compile system unconsciously.

I don't use SDL as a cross-platform input. It is too heavy. A lightweight input and output is actually a local api.Android IOS touch event,Accelerator temperature, etc.

I didn't notice blender2ogre is official,sorry.I thought that only the Tool in the ogre main repository is the official.

Yes, I find OgreAssimp. For a long time.The problem is that Assimp itself is difficult to integrate.blender2ogre is better.

imgui is good optional choice.

I know that most problems have solutions, but this adds to the cost of learning.

This actually depends on who the ogre will be as a user in the future.

As an experimental product? an introductory tutorial for beginners? or an engine for the enterprise.

[paroj]

This actually depends on who the ogre will be as a user in the future.

As an experimental product? an introductory tutorial for beginners? or an engine for the enterprise.

Thats actually a good question that our last user survey, does not answer appropriately. Will consider it for the next survey.

[Tobse]

I like Ogre. What I am missing is a way to easy contribute. For example some time ago I want to edit the wiki. How? I did not find any information.

[lonewolff]

We have prebuild sdk packages Form windows, including different language bindings. So i cannot agree Here.

Where? I am only seeing SDK's for the 1.x family.

Ogre 2.x has been around since 2012 and there has never been an SDK for it to my knowledge.

[Erostorm]

Ogre is dying cause of many reasons.

from what i've seen in 13 years of failure since 2007, the main reason is it goes in every direction, no clear goals, no roadmap of any kind, and lately project is splitted in many versions, total cancer.

second reason it is no more user friendly, long time ago you had an official SDK, installer, you just had to install and then start coding your project and it worked.
At that time it had the majority of modern shading techncis such as normal mapping, terrain, skeletal animation, etc.
Then at some time they totally stopped bringing latest technologies to the renderer and it was getting so old, someone came and started that ogre 2.0 project, its 2020 im still waiting for a 2.0 sdk, it seems the guy jumped to 2.1, 2.2, the sky is the limit, if you cant compile, its your own fault, not his.

compare this failed project with lets say Blender.org

You will understand how a great open source project failed miserably.
Developpers not interested in modern rendering technics, not interested in modern gaming, etc, etc.

Look unigine which was also a renderer, they migrated to game engine recently, even providing SDK trial to anyone.
There are also so many other renderers that popped up on github these last years, even web based projects are light years ahead of ogre.

https://threejs.org/examples/#webgl_animation_cloth

When you see that kind of stuff beeing made with javascript and having more features than ogre itself. it makes me lol.


I keep coming on this forum from time to time over the years to see if the devs wake up and start working on that sdk seriously, but it's always a deception.

When you openup any sample browser, any version, half the demos dont work, some work on opengl, some on directx, good luck to find what is what.

like somebody is going to play games on linux, its 2020 and everyone still uses windows for gaming.
They were so quick in ogre to start coding ios compatibility, in every tutorial, ios here, ios there.

Im still trying to find a real video game on apple store, haha, one that doesnt have a hundred ads forced on your screen per minute.

So that ogre project is not going anywhere.
I was hoping this patreon would make things evolve positively, but it seems that the guy doesnt understand the concept of patreon, if you dont have anything to give in return, no one is going to subscribe to the patreon, its not a donation system like paypal.

[paroj]

from what i've seen in 13 years of failure since 2007, the main reason is it goes in every direction, no clear goals, no roadmap of any kind, and lately project is splitted in many versions, total cancer.

The project split was made to make clear in which direction things are going, namely that Ogre1 and ogre-next differ in cadence and goals and, particularly, that ogre-next is not simply the most recent version of Ogre1.

Regarding Ogre1, we are not trying to compete with AAA game engines as this is beyond our resources. Instead, we try to keep the focus on what Ogre always has been; a versatile rendering solution. Here, we focus on a stable API to ease using it as a component of a larger system, while upgrading the implementation under the hood for fast rendering paths with new rendering APIs.

Speaking of integration: we now provide SDKs for Python, C#, Java and Android for each release which makes Ogre available to people outside of the C++ world, just wanting to visualize "something" quickly.
Here, the Effects like material files offer an easy interface across all supported RenderSystems.

Stability and ease of use are also the main requirements of our users according to the current survey. Therefore, this is the direction we will continue. As was already noted in this thread; Ogre is in very specific niche nowadays, so for now our goal is to defend that position.

[EricB]

Please note, I am not an Ogre developer. Although I used to provide a lot of community assistance on the IRC channel.

Ogre is dying cause of many reasons.

from what i've seen in 13 years of failure since 2007, the main reason is it goes in every direction, no clear goals, no roadmap of any kind, and lately project is splitted in many versions, total cancer.

Yesh, what a lamo.

First off, failure since 2007? It's been a long time since I updated this list, but 64 games released on Steam by Q3-2017 (before Valve opened the flood gates) is pretty damn impressive. That's roughly a half percent market share. In 2014, Ogre had a full 1% market share on Steam. And that's not including games that use all or parts Ogre, that we couldn't find.

So Ogre is doing quite well in the gaming industry for what it is. And gaming isn't even Ogre's focus.

second reason it is no more user friendly, long time ago you had an official SDK, installer, you just had to install and then start coding your project and it worked.

Folks looking for "user friendly" game engines left Ogre in 2012 for Unity, UE4, and these days Godot. If you're using Ogre3D to make games after 2012, you need a very damn good reason to be rolling your own engine.

At that time it had the majority of modern shading techncis such as normal mapping, terrain, skeletal animation, etc.

Ogre had a lot more developers looking to make game engines in those days. Game making devs have all long since left to work on premade engines. Even Sinbad.

Then at some time they totally stopped bringing latest technologies to the renderer and it was getting so old, someone came and started that ogre 2.0 project, its 2020 im still waiting for a 2.0 sdk, it seems the guy jumped to 2.1, 2.2, the sky is the limit, if you cant compile, its your own fault, not his.

If you can't compile Ogre, you shouldn't be using Ogre. That's always been the case. It's not a drag and drop game engine. It's a rendering library for those of us that don't want to write our own rendering stacks. If you want something easy for game making, you use a full blown game making engine. I've been on Ogre since 1.3, and I think as of 1.11 with C-make, it's never been easier. But that might be me.

compare this failed project with lets say Blender.org

Blender is a failed commerical product that found a FOSS niche against 3 extremely expensive products owned by the same company. The only reasons Blender did well, is because 1) hobbiest don't want to spend $2000-$3000 on 3d modeling software. 2) It's freaken hard to write a full blown 3d modeling stack. 3) And very little competition. Throw in a #4 after the hobbyist became semi-professionals, businesses found a way to cut a few grand per head... Had blender actually been a good product for most its life, I imagine Autodesk would have been in a world of hurt.

Ogre had lots of competition from all angles. FOSS competition, Game engine specific competition, Industrial competition. And then the big players, AAA high dollar engines that went with a freemium model. Imagine if 3dsm and maya were 100% free. Blender wouldn't be anywhere near what it is today.

So your comparison is bad.

You will understand how a great open source project failed miserably.

Considering Ogre is 20 years old, still actively developed, and still actively used in a vast array of products, I'm not sure how it's a failure. Is Ogre in a decline from it's peak? Sure. But it's no where near the bottom or anywhere close to a failure.

Developpers not interested in modern rendering technics, not interested in modern gaming, etc, etc.

Ogre is not a game engine. It has NOTHING to do with games. You can make a game with it, yes, but it doesn't matter if the developers are interested in games. They writing a rendering engine.

As for modern techniques, I don't keep up with development too closely. But Ogre2.x is fairly modern. Ogre 1.x is for those of us with legacy products. I do not want modern techniques in Ogre 1.x, because it would break the half million dollars worth of software I have wrote for my product. I just want incremental improvements and bug fixes. Use Ogre 2.x for modern techniques.

Look unigine which was also a renderer, they migrated to game engine recently, even providing SDK trial to anyone.
There are also so many other renderers that popped up on github these last years, even web based projects are light years ahead of ogre.

https://threejs.org/examples/#webgl_animation_cloth

I'm not seeing anything you can't do in Ogre. Although I will agree Ogre needs some easier ways to implement eye candy shaders.

I keep coming on this forum from time to time over the years to see if the devs wake up and start working on that sdk seriously, but it's always a deception.

When you openup any sample browser, any version, half the demos dont work, some work on opengl, some on directx, good luck to find what is what.

If that's a problem for you, you're more than welcome to contribute.

like somebody is going to play games on linux, its 2020 and everyone still uses windows for gaming.

Ogre is not a game engine. Stop referring to pointless gaming.

That being said, other than with my own product, I do not use Windows for gaming. Hell, I don't even use Windows on the internet. It's a stupid security risk to do so.

They were so quick in ogre to start coding ios compatibility, in every tutorial, ios here, ios there.

Im still trying to find a real video game on apple store, haha, one that doesnt have a hundred ads forced on your screen per minute.

The development team goes in the direction that they need, or are paid to go. Want them to go a different direction? Contribute or pay them.

So that ogre project is not going anywhere.

Ogre 1.x is going quite well if you ask me. I pass it along updates every now and then to my 30,000+ users. And they're happy. Eventually I plan to pass along some of those profits back to Ogre.

However, it might not be going where you want it to go.

I was hoping this patreon would make things evolve positively, but it seems that the guy doesnt understand the concept of patreon, if you dont have anything to give in return, no one is going to subscribe to the patreon, its not a donation system like paypal.

You don't have to give away anything for a patreon. I know lots of people who use it simply as a pay me so I can keep working on this stuff system.

I do however think you don't know how not to be a jackass. But I digress.

[nevarim]

Hi all

if I can have my say on the problem that ogre has is the large number of beautiful abandoned ideas that simplified the life of the Ogre user.
skyx, a great program that created spectacular games with sky.
Hidrax, a nice program that managed the sea very beautifully.
another addon whose name I managed the walkable and non-walkable paths.
all these programs have been abandoned and are no longer usable with the new versions, it would be a great thing for those who know how to do it to try to make a downloadable patch (entirely with the addon itself) in order to easily integrate them with ogre.
As for the guide, the last guide you were considering considers VS 2008, should be updated at least with the penultimate version (2017).
another big problem are the numerous old and not updated guides on the ogre installation wiki page (before getting to the one posted here I went to at least 4 different old guides)

this is my opinion: D

Neva

[paroj]

skyx, a great program that created spectacular games with sky.
Hidrax, a nice program that managed the sea very beautifully.
another addon whose name I managed the walkable and non-walkable paths.
all these programs have been abandoned and are no longer usable with the new versions, it would be a great thing for those who know how to do it to try to make a downloadable patch (entirely with the addon itself) in order to easily integrate them with ogre.

This is what I am doing at https://github.com/OGRECave

As for SkyX & HydraX specifically: their License prevents them from being included. At OGRECave I try to keep everything compatible with MIT, so you can copy & paste code freely. I also approached the author of SkyX & HydraX for re-licensing, however I got no reply.
Caelum was re-licensed as MIT though.

You can find a updated SkX here though: https://github.com/paroj/SkyX

As for the guide, the last guide you were considering considers VS 2008, should be updated at least with the penultimate version (2017).
another big problem are the numerous old and not updated guides on the ogre installation wiki page (before getting to the one posted here I went to at least 4 different old guides)

please post a reference to any invalid guide you encounter. I try to add banners on top that hilight that they are outdated and point to the new guide, which in this case is: https://ogrecave.github.io/ogre/api/latest/setup.html

[BohdanKornienko]

actually QA did improve significantly in the 1.x series. We now have unit tests that run after each commit and the visual tests are executed. (however they currently have to be verified manually)

Is not it something where automatic visual recognition system can help?
I am kinda related to Computer Vision field. It would be interesting to look at.

[paroj]

Is not it something where automatic visual recognition system can help?
I am kinda related to Computer Vision field. It would be interesting to look at.

the issue is not the comparison itself - we already compute metrics like PSNR and SSIM. It is rather doing the CI integration.

Here, we cannot simply upload the reference images somewhere and then compare against them because the rendering depends on the used driver and/ or GPU.
Instead we have to upload the rendering from the specific CI system. I did a test some time ago and it seems that the Mesa Software Render is sufficient:
https://bintray.com/ogrecave/ogre/downl ... esults.zip

Now that I have written this post: it should not be that different from the SDK we deploy already. When a commit hits master it overwrites the reference images while all pull-request are compared to the reference images beforehand..

Hmm.. I just need some spare time to do it..

[nevarim]

Caelum was re-licensed as MIT though.

caelum is compatible with ogre 2.x? i tried to compile with old ogre and wasn't possible to compile it

if u compiled it with newest ogre can u give me where u download newest version?

thanks
Neva

[sercero]

Hello,

You can download the latest Caelum version from here:
https://github.com/OGRECave/ogre-caelum

I think it is compatible with the latest OGRE3D 1.x versions.

[nevarim]

Hello,

You can download the latest Caelum version from here:
https://github.com/OGRECave/ogre-caelum

I think it is compatible with the latest OGRE3D 1.x versions.

i'll try to compile and use with 2.x

what about this?

http://wiki.ogre3d.org/tiki-index.php?p ... e%20Plugin

[paroj]

Caelum was re-licensed as MIT though.

caelum is compatible with ogre 2.x?

no

i tried to compile with old ogre and wasn't possible to compile it

it was only tested with Ogre 1.12, so this is probably what you need.

[cc9cii]

Just came across this thread and thought I might add my $0.02...

Background: I know a little about Ogre 1.x - definitely not an expert, but not a complete novice, either. I've been using Ogre since 2014. Stayed with Ogre when OpenMW switched to OSG in 2015 (I believe I am the only one still using Ogre with OpenMW).

About 3 weeks ago I decided to try porting to Ogre 2.x. It is a humbling experience. It seems like one needs to know a lot of low level details, GPU technologies, etc, to be able to do the porting. That is, to simply replicate what already works in 1.x.

One of the key reasons for me to stay with Ogre was because I didn't have to know all the detail. Ogre dealt with it. If I need some tricky stuff done, there are still useful info in wiki or other examples that can be found with a quick google search. But with 2.x, there's nowhere near the amount of info to help with the issues (hardly any, IMHO). And the API keeps changing! Which means other things break - like MyGUI, BtOgre and other custom libraries built using the old API. Without use-case examples it is difficult to know how to "un-break" stuff.

And it turns out that some things from 1.x simply won't be supported going forward. The main point about what I'm doing is to be able to use all the existing assets (meshes, textures and other resources) from the original games and from the Modding community that's been around 10+ years (thousands of MODs on Morrowind, Fallout 3/NV, Skyrim, etc). If some of these assets can no longer be used with Ogre, then either I have to become an expert and do it myself or go find something else that will. Also, if I need to become an expert in low-level stuff, then one of the reasons for staying with Ogre is no longer there.

Just to be clear - I understand that there are lack of devs/time and I appreciate all the hard work everyone is putting in, be it code updates or helping out in the forums. And I'm probably not in a good frame of mind having spent 3 frustrating weeks going nowhere, so please take what I write with a grain of salt. But to me Ogre seems to be going towards an even smaller niche.

[EricB]

Just came across this thread and thought I might add my $0.02...

Background: I know a little about Ogre 1.x - definitely not an expert, but not a complete novice, either. I've been using Ogre since 2014. Stayed with Ogre when OpenMW switched to OSG in 2015 (I believe I am the only one still using Ogre with OpenMW).

About 3 weeks ago I decided to try porting to Ogre 2.x. It is a humbling experience. It seems like one needs to know a lot of low level details, GPU technologies, etc, to be able to do the porting. That is, to simply replicate what already works in 1.x.

Ogre 2 isn't made to port old code. You would be better off using Ogre 1.12, as Paroj's goal is to move Ogre 1 code base forward, backporting new features from Ogre 2 where he can, while making it easy to maintain an Ogre 1 code base.

Use Ogre2 for new or small code based projects, and yes, it's very much a work in progress. I'm personally not a fan of them releasing it out of Beta stage without significant documentation and a fairly stable API. But it is what it is.