This post can’t be empty.
1 Like
I don’t understand much of why we are doing whatever it is we are doing here.
You have a bunch of surfaces in two sets. They are all planar:
Whenever a pairing of these planar surfaces intersects more than twice, the resulting lines will be collinear. So just pick the first line in the list.
You can also just stick the surfaces into Plane | Plane intersection, and these will give you a single line segment, assuming it is not important where on the intersection line this line segment sits.
Where is this file for us to look at?
Yes that’s what I meant here:
I had done that - my data was missing some lines and I probably over-explained something fairly simple so it should all be forgotten! 
But I do appreciate your insights, thanks for that.
Forbidden and locked away haha - I wouldn’t burden you with that mess at this point.
If I were to work off that old thing then I’d just start it differently to cancel the need for this workaround you helped me with!
This doesn’t mean I am done asking questions, though 
just give me a couple of days!
Thanks Volker!
Dear colleagues 
As promised, here we are again, stuck in sand.
If I started a new topic it’d be “make crv direction consistent for BlendCrv” or “How the heck does BlendCrv work?”
I have these ‘medial axis’ segments:
The BlendCrv command in Rhino just works! It doesn’t seem to care about anything. But in GH specific curve direction matters for the BlendCrv component?
Not that I understand how it works…
Consecutive curves facing X ish, blend works:
Consecutive curves (flipped) facing negative X ish, blend “works”:
Obviously it does ‘work’, but I wish to make a nice blend the rhino-don’t-care way from the end of one to the start of the next one.
Usually using Dot Product or Cross Product (depending on the case) handles comparing vector directions easily until you get naughty 3D ones that look like they might point in a direction but it’s an illusion once you orbit around them.
Will I have to project the lines to the ground or compare to ‘Z ish’ direction as well? Brain is blocked!
Here’s the sample file in case you have a chance to take a look:
MakeVectorsConsistent.gh (434.4 KB)
I mean, I feel you’re pretty close to a solution there with Curve Proximity…
1 Like
I wish - did you try it?
Also, curve proximity is limited, and it’s not doing anything related to connecting curves unless you make a line shared by proximity points, but that’s not a pretty blend. Alternatively, crv proximity is great for something like this to build connecting beziers or tangent crvs but those rely on crv tangent(s). For some of these curves the tangent at end points will make the connecting segments too wacky as opposed to a nice transition I can soften the ‘bulge’ factor for. Lastly, and seems to be a separate thing, it’s still odd how the blending behaves when crvs get flipped: their directions are consistent but the result wasn’t. Now that I think about it that part might just be related to the crv order - unsure.
You’re not looking to use it as a blend. You are looking to use the linear connections between the curve segments to create a polyline which will can then be referenced to correct the direction of the original curves.
Make vectors consistent VR 01.gh (425.3 KB)
(Above no polyline. I simply joined the curves with the linear connections between them with Join Curve, then pulled the curves back out.)
1 Like
Wait. Da@$m you’re right 
Let me try from here
Before I quit my job and never touch gh again, I just want to say thank you, you’re a genius. Please don’t ever explain to me how I failed to even begin to imagine this was the way.
*ps. I still don’t know how BlendCrv works but who cares.
1 Like
I hope you’re being facetious. Also, if you don’t need different bulge factors for the different ends, you can just throw the list of curves into Connect Curves.
1 Like
Haha, did you have to rub it in? Yes - I’m in happy curve world now. I promise to one day come up with something actually challenging for you to help me with! 
