Final Tile Tent



This tent structure is intended as a semi-private space for one or two people to talk quietly or operate a personal communication device without interfering too much with the atmosphere of the space it is erected in.

Working from the initial tent gesture and a hexagon based tiling pattern we developed a honey comb like structure. Each cell of the structure is capped with a six sided faceted face that combines with two of its neighbors in a rotating triangle grouping. These faces transform over the structure to create a highly complex and highly ordered geometry that takes a moment to resolve in the observer’s eye. At scale the honeycomb structure has some interesting performance properties including insulation and sound dampening.

Phone Booth / Tile / Tent



My rough plan was to build some sort of hut or tent people could enter to make phone calls or facetime calls or just be with their devices without interfering with the scholarly activity going on around them. I wanted to bring in some geometry I had worked with last semester – a sort of rotating triad of irregular hexagons.


After a lot of back and forth I ended up creating the base tile pattern using the triangle panel tool in lunch box on a simple square surface. I then split the triangles up into groups of three quads using weaverbird’s constant quads component. After some messing around with the points in each quad I had a rough 2d tiling pattern similar to what I was looking for.


Unfortunately the quads were in groups of three which were in turn in columns and I needed them to be paired along their long edge so I could treat them as hexagons.



After a lot of going down blind alleys and one desperate call for help I got them broken up into two a list of paired quads. There are still some holes to patch / irregularities to work out but the basics are there.

From there I dimensionalized and further pushed the character of the hexagons. Which resulted in this:



As you can see, the offset is causing the tiles to separate from one another. This problem was fixed later in the process. The end result sort of surprised me with how organic it looks. I’ve been comparing it to the Wild Things from Maurice Sendak’s book. No doubt the weird eye hole in the back is contributing to that vibe.

f007 f008


And some gifs showing different parameters in action:

a3 a2 a1


The offset issue is fixed. Next steps include redesigning the form finding process, debugging and figuring out how to unroll each cell for laser cutting.

Parametric Bench

tokyobenchUsing the Frank Ghery Tokyo Bench as precedent I developed a parametric bench that could be fit to any input planar curve (or at least any that I tried – there is probably a curve out there that will break it.)

The Grasshopper Definition allowed the following parameters to be manipulated:

  1. Input Curve
  2. Density of “spine” elements in each section
  3. Rib count (also number of sections)
  4. Seat Depth 1
  5. Seat Depth 2
  6. Seat Height 1
  7. Seat Height 2
  8. Position of defining profiles relative to (t) on input curve
  9. Range of lengths of protruding end spines

Due to what I am currently guessing is a data matching problem the model runs a little heavy, so I did not animate the parameters. When I have a chance to refactor the definition I will document that process here.

Five Family Members

To keep things simple I only manipulate parameters 1-7 (with the exception of #5 where I went CRAZY)









Alex Smith Bench Project

Full Bench ASJS_bench_ribsASJS_bench_skin

Possible design parameters for extension or manipulation in grasshopper.

  1. Input curve – The bench could be generated on any curve
  2. Number of ribs – or possibly density of ribs
  3. Number of spines – or density of spines
  4. Thickness of spines
  5. Seat back height at highest point
  6. Seat depth at deepest point
  7. Distance from end A to beginning of transition to seat
  8. Distance from end B to beginning of transition to seat