The Mosaic Dish Scrubber: A Not So 2D Design

This Is my 2D surface design. it is to be put simply, a mosaic tile pattern forced into a wave. for whatever reason I thought this was a great idea, I wanted to know how to create this shape and I really just decided to figure it out regardless of how difficult it was. If the wave isn't super obvious it will make more sense from this angle.

but to make this texture, I started really with just circles. Repeating and trimming them until I had a shape like this. 

Afterwards I just started splitting and duplicating until I had enough lines to make each shape into its own closed curve that could be extruded individually of one-another. once i had that finished, I arrayed the shape 20 times in the x direction and 20 times in the Y direction. 

Then (and this may be the only command I've used that we haven't covered) I placed a series of points around the outside of this 20X20 square. Once I had them, I elevated each one a slightly different amount, and then drew curves between the points. 

This operation gave me the edges that would define my wave. and then by selecting all surfaces and entering "edgesrf" I created this sheet out of non-planar lines.

Well then. Now I had my geometry and my obscure surface, and from this point on, things went horribly. My initial idea was to 'project' the geometry upwards onto the surface, then extrude individually the different shapes from the sheet. The problem with that was that after I projected my geometry, I had no way of individually selecting different the source shapes without individually clicking on each one. also, because of all the shapes being coincident on top of one another, each selection took about 3 clicks on average. To put that into reference, the least plentiful of the 3 shapes in the design had 400 identical siblings. and after comleting my first 1200 clicks I realised that this plan was slightly unrealistic. At this point I also realised that after I had all of my extrusions would be impossible to cap, because each of the the openings was strangely shaped and non-planar.

After talking with Jin and Bryan, I gained a couple insights. 1. I didn't need to project at all, because I could just extrude between two sheets 2. If I extruded through sheets the resulting split would also give me both of my "caps" if you will and 3. and this really is a big turning point. If I put each individual shape into its own layer, then when they are all duplicated 20 fold, I can simply deselect that layer when I want to only select one shape to extrude. this means that my previous 1200 clicks just became one effectively. this however meant that I needed to return to the beginning. I redrew all of my original geometry, but instead of being all boring and black, it was now colour coded by layer.

But even with colour coding and the other suggestions, my model still wasn't performing properly. It was nearly killing the computer I was on, and Rhino was also giving me a plethora of error messages. It really didn't click, until I went back to the geometry I started with. When I zoomed in incredibly close to each intersection, there were errors. things either didn't line up, or they overlapped in undesirable ways. These faults were being multiplied thousands of times across the sheet of geometry, and then when I pushed it to 3D, Rhino simply couldn't fix it. So I again redrew the model and this time extruded properly and looked something like this:

But after I split the sheets that ran through the centre of the model, they became thousands of tiny pieces again, and I couldn't use the layer trick because they were all from the same sheet(and therefore layer). This meant that I was back to having the same issue of having to select thousands of tiny tiny shapes. 

The Resolution

After redrawing it again, some concessions were made. I came up with a way to keep the sheet solid even after all of the splits. Instead of having my lines come to exact points, they would all have small channels running in between them, like so:

this is the only way that the model became successful. after I made this change the 3D commands ran reasonably (they were really slow but the program didn't crash). and the end result was still what I hoped for. I would like to point out that this model still isn't perfect. There are still unbound surfaces in the middle of it, and I have no Idea why. But anyways thanks for reading this, I hope it wasn't too painful. 

supporting images: