Would it be possible to add the ability to select and hide specific wires? Also the ability to make instances of nodes would also be great, similar to how you can make instances of objects in the 3d view. This would allow to instance the nodes you need access to regularly and organise them separate from the node tree in a logical layout like below. Tweaking the instances parameters would then also modify the original node which is located in the node tree :
The instanced nodes should have no inputs or outputs so they cant be wired to anything.
maybe @jacqueslucke you could consider this for your everything nodes project? Although it’s mainly the compositor I need it in most, it would probably be even more useful for complex particle, mesh editing and animation node trees. Perhaps the area of instanced nodes could act as a front end which could be opened up to show the full node graph, like a HDA in Houdini. This would make it really easy to share complex set ups where the user only has to worry about tweaking the front end instanced nodes. Sort of like a group node but with a front end GUI which exposes a selection of full nodes which a user is likely to need to tweak. Save them having to try and work out what does what in a complex web of wires. You could nest them too to seperate nodes specific to a certain task into more logical blocks for added readability.
yep, that’s the only way to work currently in blender if you want the graph to be tidy, but it’s very laborious, and can be a bit of a pain if you want to modify it afterwards. I’m not suggesting to do away with wires though, just the ability to hide them so you can layout visually with layout frames, and then maybe use the annotation tool to draw lines to indicate data flow, if annotation tool is available in the compositor that is, i’ll have to check.
I spent a couple of years in Houdini, and they have a great system of generally only one wire between nodes, which carries all of the information from the first node to the second. Then instead of having sockets for each parameter, you have either a dropdown, or you type in the name of the parameters you want from the flow of data that’s running through the single wire. clicking the parameter name flicks between showing the value and the data name being referenced. The node graphs are MASSIVELY more concise, and the data flow is much easier to understand, and more importantly there’s no need for all of that donkey work you’ve done in the example you posted above.