Makes sense.
How will you decide which wavy shapes are good and which are bad?
I’ll try to take a look this afternoon (super late here in Stockholm).
Mind if I ask which university? I like to keep an eye out for ones I can recommend to people interested in this type of stuff. 
feel free to DM me if you’d like to not put that out there.
I was thinking there should be a way to re-run the process again and again until I get a combination of claddings I like. Otherwise I will adjust the parameters and narrow it down.
We are neighbours then haha. I’m in Copenhagen, the school is Københavns Erhvervsakademi. The course is called Architectural Technology & Construction Management. And tack så´mycket !
So each of these is a single panel, and you’re going to generate each uniquely and then try to assemble them into a unified design?
It seems you’d be better off taking the face of the surface and segmenting it with waving lines, as this would ensure a cohesive result, but I’ll think on it when I take a look later.
thats really great i wish i have that option when i was at that school, but in that time we dont even have computer first start up with that in last semester with acad R 12 in DOS;) …hahahah ;)n ha en god dag
I have it setup like this at the moment:
However, even if I change the distance to place the different solids at different points on the X axis, all of them overlap at 0. Why is that?
A - make these random… Do them from a range and output 3.
Feed into just 1 of your “Facade” nodes.
B - Make this a range as I showed you earlier.
Feed this into just 1 Geometry Translate node.
Any chance you have a sketch showing what the end result should look like?
Something like this. There will be a slider controlling the number of copies, in this case three. But I think what I posted above may be a better solution, as I can individually control the variation of each solid.
Sorry it took so long to get back to this - lost the thread as the title wasn’t very clear (edited it accordingly).
The random is already occurring inside the custom node by way of a series of controlled Math.Random nodes, which means there is no need for any randomization in the workspace environment. The range works fine.
That said, there is no way to control random - you only ever get random, and the result you get today will never be attained a second time. As such if you run, get something almost right, edit to make things a bit better and then run again you may be further than you wanted to start with. As such moving the random into a controlled variable is likely preferable in a non-educational environment. But since we are learning here, letting random play isn’t a bad thing.
As such leaving the panels as they are and sequentially shifting their width is likely the best bet.
Thanks a lot! I will definitely be looking into this. However I changed the design slightly (typical architecture stuff). I will be sharing my final result here .
