![](https://attach.openclonk.org/avatars/30-1442.png)
![](https://attach.openclonk.org/avatars/65-0196.png)
![](https://attach.openclonk.org/avatars/31-31746.png)
What was the problem with the backend again? This is a bit frustrating, because it is essentially the same situation as two months ago (or 20 months, if you count the time since the screenshot was posted)?
![](https://attach.openclonk.org/avatars/65-0196.png)
However, I have no experience with the framework that was used, so I can't really help at all. My take would be to re-write it in Python and Flask; but I don't want to do that alone either.
I have no experience in either of that, but my best guess is: Even if we re-write we may be closer to an actual implementation than if we just wait for the existing backend to get finished. After all, if there is no real progress for 2 years it is likely to stay that way.
Started something similar last summer. (Which I'll continue this summer.)
Although it's based on JSON rather than XML, adapting the ingame UI isn't a problem - I already did that for LC (which still fails due to some weirdness in StdScheduler).
Although it's based on JSON rather than XML, adapting the ingame UI isn't a problem - I already did that for LC (which still fails due to some weirdness in StdScheduler).
I reimplemented a (slightly different) interface because I didn't understand Larry's code at all and wasn't quite contented with some details I forgot again...I think it was something with the upload output. I don't know about another Larry "implementation" though.
(And apparently I'm not the only one failing to understand Larry's source...)
(And apparently I'm not the only one failing to understand Larry's source...)
The original was lorry, then there was larry, and yours is parry. It is a bit amusing to have so many versions of it.
One question about Larry then: Is the server that it connects to fixed? If not, could we live with providing the GUI already and just say "there is nothing to download yet".
It's set in Quasar's options, so you'd need to stop, recompile and start the GUI's server instance.
![](https://attach.openclonk.org/avatars/65-0196.png)
Provide incentive to finish it? I do not know about the way it works, but if you could connect to a configurable master server, then adding the GUI is a separate thing from getting the server running. Of course, adding the GUI is just confusing newcomers then.
![](https://attach.openclonk.org/avatars/31-31746.png)
Still there. Recently I did engine contributions, but moved to other projects, because I lacked feedback.
Also, I am constantly working on more or less transforming concrete implementations into more abstract and modular concepts, so that they can be reused easily. This happens outside of the game repository, though, especially, because I do not want to break the game. Do we want to integrate those somehow, or leave it separate? For example, the AimManager defines things explicitely for our game (walk speed, animation replacements, and so forth), and ranged weapons use these in the form of an animation set.
The abstracted version is a stance transition system, where each stance can define behaviours as it sees fit. A behaviour then works simply applies and resets when the stance is active/deactivated. This works quite well in my test scenarios, but adding it to the actual game means A) changes in the existing system and B) breaking projects of other developers. That was one of the reasons why I thought about a "from scratch" approach to either the game content or the engine a few months ago.
The abstracted version is a stance transition system, where each stance can define behaviours as it sees fit. A behaviour then works simply applies and resets when the stance is active/deactivated. This works quite well in my test scenarios, but adding it to the actual game means A) changes in the existing system and B) breaking projects of other developers. That was one of the reasons why I thought about a "from scratch" approach to either the game content or the engine a few months ago.
![](https://attach.openclonk.org/avatars/65-0196.png)
>Do we want to integrate those somehow, or leave it separate?
Using more abstraction has some pro and cons. Pros are that they are forward-compatible and it might be easier to implement yet-unknown changes to the game contents and might or might not lead to more modular and easier to understand code. Cons are that they are usually more time consuming to write and sometimes harder to use.
Especially if it requires the scripter to fully understand the animation system (if they e.g. want to add a new weapon), it might be a turn off.
Soo, I guess that question can't really be answered without a specific pro/con discussion. I am not the biggest fan of the animation stuff/aim manager because it led to hard-to-fix bugs in the past (e.g. being stuck in the drawn-bow-animation). But that also might be an indicator of how challenging it is to write a good system that is robust against all the things that can happen (same as the power system).
I agree with your first point. On the other hand, the animation system and aim manager already require someone to fully understand how to implement it (including that the AimMode is redundant, since AIM_POSITION works only with AimAnimation and AIM_WEIGHT works only when AimAnimation and AimAnimation2 are set, and so on. Also, I was quite surprised to find out that the WalkSpeed and WalkBack values are absolute, and not just a relative value for scaling the original speed). It is also quite intertwined with the hand action / inventory stuff, which is ALSO quite hard to understand, and neither of those systems are abstract, but specialized for this game. This is more a general rant than an actual pro/con discussion now :)
![](https://attach.openclonk.org/avatars/30-1442.png)
![](https://attach.openclonk.org/avatars/30-1442.png)
Some of them were bug fixes that I needed feedback for, but that could be handled in the IRC channel quite well. I was talking about the movement code - you were the only one to test it and give feedback. Now it is two months later and I just put this up as a pull request.
I'm pretty much an observer.
In better news, Isilkor seems to be working on a better font renderer. https://i.imgur.com/CXHcRHv.png
In better news, Isilkor seems to be working on a better font renderer. https://i.imgur.com/CXHcRHv.png
![](https://attach.openclonk.org/avatars/29-2030.png)
![](https://attach.openclonk.org/avatars/30-1442.png)
When your retirement comes, Clonk will have to be nothing short of a fully immersive VR experience.
![](https://attach.openclonk.org/avatars/46-1025.png)
![](https://attach.openclonk.org/avatars/59-2177.png)
Hi all, just wanted to make a little note here that there are people not involved in development but still frequently having an eye on the project looking forward for new things to see and of course are willing to test stuff!
![](https://attach.openclonk.org/avatars/572-36568.png)
But, really guys the game is looking better than ever. Don't give up when the game is getting good! Also, I would love to dive into learning how to develop content, but I feel there should be an all in one editor for OC and some tutorials would be a good idea. Don't make your devs have to go digging around to put all the dev tools together and figure stuff out. Make development easy take cues from CR and the original editors ease of use. I love how the editor was linked (packaged with) the main game in CR. I would suggest at least make a package that ships the tools and the game together, even if you keep the ability to download them separately.
Anyway, that's my feedback for now! Big fan!
![](https://attach.openclonk.org/avatars/65-0196.png)
The only sad thing is that there currently isn't an all-in-one editor like in CR and even though people have tried creating replacements, they were never really finished :(
How polished are windmill and nanowindmill? Whats the difference? Are the OC code extensions still linked to the abomination that is Eclipse? I would consider dev work if the tools weren't so big. By the way, how do you get models from blender to OC? I think, that is the really issue for the game progress right now, the game is getting more and more awesome to play, but learning the dev tools is a bit of a pain.
Correct me where I'm wrong, because the reality is that I haven't tried OC dev work in a long time (I didn't know what I was doing...), and back than I thought the only way was Eclipse. I couldn't figure out to get models over to OC. I felt like it was some demented thing like exporting from Blender to Ogre to OpenClonk.
Contrast that with CR where I could literally code with a text editor and make content with a single png image. Ease of development...
And where is everyone? What happened to Newton? Who is organizing our wonderful OC team? I should dig around the source code and get familiar with everything, because I want to contribute to my favorite game...
Correct me where I'm wrong, because the reality is that I haven't tried OC dev work in a long time (I didn't know what I was doing...), and back than I thought the only way was Eclipse. I couldn't figure out to get models over to OC. I felt like it was some demented thing like exporting from Blender to Ogre to OpenClonk.
Contrast that with CR where I could literally code with a text editor and make content with a single png image. Ease of development...
And where is everyone? What happened to Newton? Who is organizing our wonderful OC team? I should dig around the source code and get familiar with everything, because I want to contribute to my favorite game...
![](https://attach.openclonk.org/avatars/30-1442.png)
> Are the OC code extensions still linked to the abomination that is Eclipse?
OC was never linked to Eclipse. I personally never coded in Eclipse. OC always came with the c4group.exe to unpack all game content into explorer-readable files that let you edit them with whatever text editor you wanted.
> By the way, how do you get models from blender to OC?
You had to use the Blender exporter plugin made by Randrian (I am not sure where to download it nowadays). The exporter would generate the .mesh and .material files needed. But there are some things to watch out for as they will cause you trouble.
The exporter is by now outdated and doesn't work with the newest version of Blender. Exactly what Randrian said here.
> And where is everyone?
Moved on to other projects. The core dev team still knows each other in real life. Newton is working on StreetComplete (with no intention of coming back to OC). Apart from that there is no active development except what Marky is doing in his repository. OC is inactive right now with little chances of it ever reviving.
![](https://attach.openclonk.org/avatars/572-36568.png)
Is there any other way to get meshes into OC or are we stuck with an exporter for an outdated version of Blender?
How did the guys from Codename Modern Combat (Marky right?) get new meshes into the game or are they using the outdated exporter?
I had forgotten about the c4group.exe. Thanks for the reminder!
The dev tools aren't as bad off as I thought, if I can set up a functional system for making more meshes and materials...
Anyways, thanks for the feedback!
![](https://attach.openclonk.org/avatars/30-1442.png)
![](https://attach.openclonk.org/avatars/572-36568.png)
What kind of pitfalls are we talking about? Is it related to exporting blender's bone rigging or what kind of things should I be aware of, if I want to work on animated meshes with bones?
![](https://attach.openclonk.org/avatars/30-1442.png)
In the end, the whole project in Blender must be one single mesh (so join everything when exporting). Also, all transforms (scale, rotation, ...) must be cleared/applied. This applies to non-animated modells as well.
Animations must be converted into NLA strips, as described here: https://wiki.openclonk.org/w/Exporting_Blender_Models#Exporting_Animations
Randrian could make this clearer, but I think there are some things that have changed since.
The exporter from Randrian does this automatically
If I remember correctly you don't need to use NLA strips anymore, especially when using an Action.txt describing how the animations should be exported. (Which is a feature Randrian added)
>Also, all transforms (scale, rotation, ...) must be cleared/applied.
The exporter from Randrian does this automatically
>Animations must be converted into NLA strips, as described here: https://wiki.openclonk.org/w/Exporting_Blender_Models#Exporting_Animations
If I remember correctly you don't need to use NLA strips anymore, especially when using an Action.txt describing how the animations should be exported. (Which is a feature Randrian added)
![](https://attach.openclonk.org/avatars/572-36568.png)
Also, If we really want to see more dev influx, send some people over Blender's forms and mention that there are opportunities to get experience while volunteering help on a project. But, we really need some introductory level tutorials on getting models from Blender to OpenClonk. (Just throwing out ideas...)
![](https://attach.openclonk.org/avatars/30-1442.png)
![](https://attach.openclonk.org/avatars/59-2177.png)
Powered by mwForum 2.29.7 © 1999-2015 Markus Wichitill