Luke
@lukedigi.bsky.social
math/cs undergrad, future software engineer
Not doing a good job guessing rn
May 6, 2025 at 2:36 PM
Not doing a good job guessing rn
Will never cease to amaze me how humans went from smashing rocks to make little blades to smashing rocks to make literal magic.
December 7, 2024 at 3:47 AM
Will never cease to amaze me how humans went from smashing rocks to make little blades to smashing rocks to make literal magic.
Back to learning OCaml today, I found out there is a difference between `fun` and `function`, where `function` allows you to directly pattern match an input, but this only works for single argument functions.
November 29, 2024 at 7:40 PM
Back to learning OCaml today, I found out there is a difference between `fun` and `function`, where `function` allows you to directly pattern match an input, but this only works for single argument functions.
On the campus of NYU today and noticed panic buttons in their bathroom stalls. Definitely a first for me
November 26, 2024 at 7:36 PM
On the campus of NYU today and noticed panic buttons in their bathroom stalls. Definitely a first for me
Today I learned: The Euler characteristic (χ = V - E + F for convex polyhedra) can similarly be calculated for arbitrary topological shapes as an alternating sum of Betti numbers and this value is topologically invariant (e.g. for convex polyhedra χ=2, for the torus χ=0)
November 19, 2024 at 4:41 AM
Today I learned: The Euler characteristic (χ = V - E + F for convex polyhedra) can similarly be calculated for arbitrary topological shapes as an alternating sum of Betti numbers and this value is topologically invariant (e.g. for convex polyhedra χ=2, for the torus χ=0)
Slightly simpler OCaml version by removing the List.fold_left:
November 16, 2024 at 10:40 PM
Slightly simpler OCaml version by removing the List.fold_left:
Continuing with rewiring my imperative programming brain I decided to write a progressive tax calculator in both Python and OCaml. The procedural version (in Python) is much easier to understand imo, but this may just be because I'm not proficient in FP yet to write good code.
November 16, 2024 at 10:34 PM
Continuing with rewiring my imperative programming brain I decided to write a progressive tax calculator in both Python and OCaml. The procedural version (in Python) is much easier to understand imo, but this may just be because I'm not proficient in FP yet to write good code.
“Soaked in wine” is going to be my new term for being drunk.
November 16, 2024 at 6:30 AM
“Soaked in wine” is going to be my new term for being drunk.
Today I learned: OCaml has a built-in optional type to return Some value or None. Pretty basic, but I'm brand new to OCaml.
I've always preferred not to use try/catch workflows - they are so non-idiomatic to me - but it works well here. Rewriting to exploit options is an eyesore (and doesn't work).
I've always preferred not to use try/catch workflows - they are so non-idiomatic to me - but it works well here. Rewriting to exploit options is an eyesore (and doesn't work).
November 16, 2024 at 4:09 AM
Today I learned: OCaml has a built-in optional type to return Some value or None. Pretty basic, but I'm brand new to OCaml.
I've always preferred not to use try/catch workflows - they are so non-idiomatic to me - but it works well here. Rewriting to exploit options is an eyesore (and doesn't work).
I've always preferred not to use try/catch workflows - they are so non-idiomatic to me - but it works well here. Rewriting to exploit options is an eyesore (and doesn't work).
Today I learned: any 2D map of regions can be colored such that adjacent regions are different using no more than 4 colors. This can proved by showing the chromatic number of the dual graph of any planar graph is at most 4. The proof handles so many cases that it’s infeasible without software!
November 15, 2024 at 4:25 AM
Today I learned: any 2D map of regions can be colored such that adjacent regions are different using no more than 4 colors. This can proved by showing the chromatic number of the dual graph of any planar graph is at most 4. The proof handles so many cases that it’s infeasible without software!
