Contact · Paradiso, Switzerland
Tell us the problem.
Finance, artificial intelligence, quantum computing, business, software or design: we take problems from any field. Describe yours in your own words; turning them into variables and constraints is our part of the work.
t = np.linspace(0, 2*π, 2000)
x, y = np.sin(3*t + δ), np.sin(2*t)
# 3 : 2 closes on itself, whatever δ is- δ, the phase: write below and it turns
- 7 crossings: 2·3·2 − 3 − 2
- 3 across
- 2 up and down
- the scope up there draws it too
- 3 : 2 is a perfect fifth, if you listen
- Lissajous, 1857: tuning forks and mirrors
- 3 : 2 is rational, so it closes at t = 2π; irrational, it would fill the square
- 22 characters in, δ = 3π/4: it folds flat
- Bowditch drew them first, in 1815: 42 years earlier
The brief
Write it down as you see it.
Plain words are enough.
Confidential details can wait. Describe the problem in general terms first; the details can follow once we have talked.
- Given
- Find
- Subject to
- our favourite kind
- read by people, answered by email
After you write
What happens next.
Three steps, in this order. Nothing is agreed until the problem is written down.
d, heap = {s: 0}, [(0, s)]
while heap:
du, u = heappop(heap)
for v, w in graph[u]:
if du + w < d.get(v, inf):
d[v] = du + w; heappush(heap, (d[v], v))We read it. We answer by email. The first answer is usually a list of questions.
We write it down. Together we state the problem precisely: what is known, what is assumed and what a good answer looks like.
We propose the work. If we can solve it well, we propose an approach and a timeframe. If it is not for us, we say so.
- s: your brief
- t: the work
- where you are today
- Dijkstra, 1956: no pencil, no paper. We use both
- ink: the tree of shortest paths from s; faint: the edges no shortest path uses
- weights are lengths, never negative, so a settled distance is final
- the paper, 1959: three pages long
We read it. We answer by email. The first answer is usually a list of questions.
We write it down. Together we state the problem precisely: what is known, what is assumed and what a good answer looks like.
We propose the work. If we can solve it well, we propose an approach and a timeframe. If it is not for us, we say so.
Correspondence
- or simply write
- 45°59′ N, 8°57′ E