Hi,
Can two (or more) GrassHopper Definitions communicate in realtime? (in the same Rhino session)
// Rolf
Not without plugins. You can bake geometry into the 3dm file using a Geometry Cache object, and then load that same geometry back into a tree structure using another Geometry Cache object set to the same name. But it requires a manual click to both update the shapes and reimport them. And it doesnāt work for non-geometric data.
2 Likes
Oh wait, you actually can use a Panel to stream textual data and a Read File component. That at least would be real-time, but it only works for textual data.
The Rhino Python sticky dictionary is shared across all GHPython instances and the editpythonscript editor within the same Rhino session, enabling one to send any data type between multiple Grasshopper canvases (and Rhino).
1 Like
I saw Mateusz ran several definitions at CITA in the background. I think it was something with the bendy wall project.
Do you know what was the workflow behind?
This maybe? Pretty sure there were no stickies involved here 
Yup.
Since this is something I was planning anyway for GH2, I wrote a very simplistic implementation using standard components for GH1 in Rhino 6.0
(ignore the audio, didnāt realise the screen capture software was also recording Harmontown that was playing on my other screen).
7 Likes
Cool!
But that wouldnāt work in GH 0.9?
// Rolf
Iām not even supposed to be messing with Rhino 6.0 for non-critical stuff, let alone Rhino5.
2 Likes
What would be the benefit of such workflow or reasons to use it?
Does it result in faster execution time or it would be more āclearerā to read the gh definition??
Let me give you and example:
I work a lot with very ādirtyā meshes. They need some very heavy processing for cleaning them up before I can start doing my atual analyse of the. If the GH definition is to recalculate starting from this cleaning process, every change (like iterative form finding) would take forever. I canāt even have the Cleaning Definition opened because it slows everything else down. For other steps in the process I have different definitions for different meshes (depending on quality), so therefore it makes sense to modularize the definition for only those steps that need variants. And so on. But at times several such steps must be active at the same time in a long process chain (due to user interaction, one can not save part-results temporarily only to open next definition, and so on, it must be one responsive chain of definitions).
Although there may be alternative solutions, like DataDams, and logical āprocess pathsā, too big definitions isnāt easy to maintain. Definitions should be like components (chained into networks), only on a higher level.
// Rolf
1 Like
In addition to @RILās example, hereās what I imagine this is useful for:
- You can split up large files into consecutive smaller files. Making each file more readable, more re-usable, and more debuggable.
- You can store various states of data to work on, and switch between them either through the GH UI or using the file viewer in your OS. You might just want different states, or you may want heavily reduced versions of the same states so that you can work on a small part of the logic and only switch to the full-blown-it-takes-20-minutes-to-compute finished model at the very end.
- It allows you to work together with other team members more easily. Person A is in charge of facade shapes, person B is doing the rationalisation, person C is populating facade elements, person D is all about creating shop drawings, machining instructions and material bills, each working in their own gh files but sharing data.
There are other reasons, but these I think are the most important.
Eventually for a proper implementation Iām imagining that both cluster hooks and these data exchanges are the same sort of object. It has to work both to and from separate *ghdata files, and also directly into other *.gh files, and of course across cluster boundaries.
9 Likes
Iāve been asking something like this before on the old forum. There is some definition which should be defined step by step, and it is perfect to have a way to do this spending time for the calculating only current step.
Thank you, it is good long waited component 
I use Geometry Cache now the most problem with it is that you have to find the one to reload in a quite a big definition and then another one etc. Adding to it boolean parameter entrance would be not bad.
Lately our files in the office have become bigger in size and harder to handle (even for a powerful iMac). Something like this would bring so much benefit to the whole workflow! Canāt wait to see it happen!
Thanks
I forgot to point out that we work in OSX environment, so no R6 for us yet. But iām positive that improvements and compatibility will surely come soon.
Thanks for the heads up though, we might try to find a windows pc and give it a go!
Isnāt this what elefront plug in does? I think it enabled something similar to this functionality.
Edit: Nvm, its not exactly what you are looking for.
Havenāt tested it extensively, but for simultaneously open grasshopper files, it looks like Link might be it?
http://www.hksinc.com/hksline/2015/01/19/linking-grasshopper-files/
David,
I am finding the Data Input/Output quite a useful addition. Thanks for that.
I am wondering if it would be possible to implement it, such as that the file path input and output can be set by using a path string as supposed to manually doing it by right clicking the component itself.