Planes on Curves

Question

My blend works, but when I wanted to respond to this post "Geometry Node: How would I convert a series of Curve Lines into a series of Planes", I proceeded using an empirical approach.

First I don't understand why the even faces and the odd faces are separated? So a part of this blend is still very strange to me. For what I know, when the number "Count" is odd the first face off the odd faces go directly from the start of the curve to the end. The second line of nodes (just under) is here to fix this problem. But as I tested lot of numbers, finally I came up with a solution, that doesn't mean that I understand how it works... I wish you can explain to me what I have done here:

And I have a second question is it possible to reorder the faces?

Thanks for your help. Here is the blend:

Edit:

After the answer of Joel Reid, I will try to simplify and narrow down this question to a more simple problem and a more simple blend. In fact my problem is to understand why those nodes produce only the evens faces and not the odds ones.

I hope this formulation is more likely to be understood, I am perfectly conscious that is a very different problem, and I am sorry for that. Thank you for your help.

Answer 1

What?

We can examine the setup and perhaps gain some insight into what it is "trying" to accomplish.

In your Odd Face Index Calculation frame, let's follow the data left-to-right… …and interrogate each in turn for "what is this node's output, conceptually?"

1. Truncated Modulo
   This node uses the [Bezier Segment's, Curve's Points, Point Cloud's] Point Count (hereafter: "c"). The node's output Value is c mod 2.
2. Switch
   A 0 or 1 subtrahend determined by c's oddness.
3. Add
   The greatest odd int ≤ c
4. Subtract
   The second-greatest odd int ≤ c

5. Add
   Pardon my syntax from here forward, where curly braces indicate the two possible values.
   {is c even? → 2c - 4,
    is c odd? →  2c - 2}
6. Index
   "Points of Reference" Curve To Points, point index field
7. Subtract
   index - 2c - {4,2}
8. Multiply
   2c
9. Subtract
   index - 2c - {6,4}
10. Subtract
    2c - 4
11. Greater Than or Equal
    bool answering, is 9's "index - 2c - {6,4}" ≥ 2 ?
12. Add
    index - {10,8}

13. Switch

combined case	Output value	simplified
c even and index - 2c - 6 ≥ 2	index - 2c - 6
c odd and index - 2c - 4 ≥ 2	index - 2c - 4
c even and index - 2c - 6 < 2	index - 2c - 6 + 2c - 4	index - 10
c odd and index - 2c - 4 < 2	index - 2c - 4 + 2c - 4	index - 8

Why?

That node tree is many things; "working", reducible, intractable, baffling…

The above was all easy enough to work out. But then what does, say, "the index 6 less than double the input count of the current index" across the realized curve's points field mean to you? You're the user of this node arrangement.

I love Geometry Nodes, and love messing about with them as much as the next Blender-er, but I humbly submit that this method of throwing nodes at a problem is a horrible way to work and clearly no way to learn. If you're borrowing and adapting someone else's complex method of accomplishing something, you should…
go back and study what exactly they're accomplishing,
borrow the parts you understand,
adapt it only with new methods and nodes and workflows you also understand.

I hope the above breakdowns perhaps gave you some insight you can use to push forward. I think this is all that can be done with your question as asked, as Blender Stack Exchange is not for requesting tutorials and project autopsies, but for pointed technical questions and solving of clear problems—which is frankly not a fitting description for the situation you're in. Please ask another question if you reduce your tree to a clear technical question. Good luck.

Answer 2

The answer was so evident that’s I am ashamed (as I should be), the first face is fixed on the 0, 1, 2 and 3 points then the second plan one 4, 5, 6, 7 points and so on… So there is no plan between the points 2, 3, 4 ,5: there is no odd faces. With just another Sample Index plug on indexes plus 2, it makes the good indexes for the odd faces.

I am very sorry for the time you lost on this question. Thank you.