Is this project abandoned?

[leissler]

Is this project abandoned?

[emrys90]

Unless something I did is affecting it, I believe all the graphs do some work on compiling/entering play mode. Which gets worse for the number of graphs in the project.

[emrys90]

I really hope its not. My project MAJORLY depends on this asset.

[rhys-vdw]

It clearly is. But it's open source so it's not the end of the world.

[emrys90]

Does anyone know of any forks of this asset being maintained? Has anyone taken over working on this?

[andybak]

Does anyone know of any forks of this asset being maintained? Has anyone taken over working on this?

For any repo you can go to Insights > Network to view forks and branches: https://github.com/alelievr/NodeGraphProcessor/network

Of course this won't catch situations where someone has just reuploaded the code into a new repo but you can try picking a class name or variable name that is likely to be unique to this repo and then doing a code search.

[emrys90]

So according to that, no one is maintaining a fork of this project?

[rhys-vdw]

AFAIK there is no community fork. It would be nice.

[leissler]

Well, at this point in time it neither makes sense for anything nor to put further work into this.
Graph Tools Foundation is well on its way and will be built into the engine from 2023 (yes, built-in, not package!).
It is highly flexible and beats GraphView architecture-wise. All graph tools (VFX Graph, Visual Scripting, Shader Graph) will run on GTF, so it's definitely the way to go. I would recommend starting to work with the pre-release to get used to the underlying concepts.

[emrys90]

Switching to something else isn't an easy option for me. My system highly depends on Node Graph Processor for its architecture, and I have close to 160 graphs made with it. Unless I'm overlooking some easy conversion of the graphs from Node Graph Processor to GTF, it sounds like a massive amount of work.

[rhys-vdw]

Do you need to change?

[emrys90]

This is a collectible card game, it could still easily be in development 10 years from now working on expansions. I need a plugin that is maintained with bug fixes and Unity upgrades.

[rhys-vdw]

I've been considering checking out Flow Canvas but we've only got 30 or so graphs so if it's really solid it could be worth the effort. Looks really good though. I have found a few little bugs in NGP, but I'm also a little optimistic about fixing bugs in my own fork. Got one gnarly bug at the moment that I haven't been able to debug, if it comes from NGP then I'll probably switch.

[emrys90]

I have a couple minor bugs with NGP that aren't a big enough issue for me to want to switch. I'm more concerned if future Unity versions introduce major changes that requires a lot of work on my end to get NGP compatible with it.

My bigger concern though is performance. NGP has a huge deserialize performance hit that gets worse with each new graph created. It's already pretty bad with a long delay entering playmode each time, and after compiling changes, and its only going to get worse as I add in hundreds of more graphs over time.

[rhys-vdw]

Interesting! I probably don't have enough graphs to cop that yet. Do you need to load all the graphs together? (btw this conversation could be moved to the discord if you prefer https://discord.gg/pEkdyTjruu)

[OneManMonkeySquad]

Switching to something else isn't an easy option for me. My system highly depends on Node Graph Processor for its architecture, and I have close to 160 graphs made with it. Unless I'm overlooking some easy conversion of the graphs from Node Graph Processor to GTF, it sounds like a massive amount of work.

Welcome to Unity... I love Unity to death but this issue keep coming up. Releasing something then abandoning it shortly after with no migration path.