Estrella
@estrella.astrallexicon.com
bio fail laugh at this user
20++ SEA they/any
fake fan of math, cs, music, anime (sometimes), games (usually rhythm but some others too)
20++ SEA they/any
fake fan of math, cs, music, anime (sometimes), games (usually rhythm but some others too)
also i can't quite figure out if its np-complete... first instinct is a reduction from knapsack but the reduction is only pseudo-polynomial. ideally i'd like to not resort to portal shenanigans but i might have to
January 14, 2026 at 11:20 PM
also i can't quite figure out if its np-complete... first instinct is a reduction from knapsack but the reduction is only pseudo-polynomial. ideally i'd like to not resort to portal shenanigans but i might have to
one implication of this is that if you have a 'range aggregate + single update' data structure (such as segment trees as mentioned by someone else on this thread), you can easily derive 'single query + range update' by storing the underlying sequence's adjacent differences instead!
January 14, 2026 at 11:16 PM
one implication of this is that if you have a 'range aggregate + single update' data structure (such as segment trees as mentioned by someone else on this thread), you can easily derive 'single query + range update' by storing the underlying sequence's adjacent differences instead!
dually, 'apply a group operation to a window [i, j]' can be thought of as 'apply a group operation to i, and its inverse to j' on its adjacent differences (the inverse of prefix sums)
which means 'apply many such operations-on-windows' can be done in O(n) time, in constant time per operation!
which means 'apply many such operations-on-windows' can be done in O(n) time, in constant time per operation!
January 14, 2026 at 11:16 PM
dually, 'apply a group operation to a window [i, j]' can be thought of as 'apply a group operation to i, and its inverse to j' on its adjacent differences (the inverse of prefix sums)
which means 'apply many such operations-on-windows' can be done in O(n) time, in constant time per operation!
which means 'apply many such operations-on-windows' can be done in O(n) time, in constant time per operation!
okay it was just a check-in. now time to lock in for the actual interview…
January 13, 2026 at 12:26 PM
okay it was just a check-in. now time to lock in for the actual interview…
SITUATION: There are 15 competing monad tutorials
January 6, 2026 at 1:20 AM
SITUATION: There are 15 competing monad tutorials
not even hardcore-functional-land, even just basics like enums and list processing
like, for all of rust's faults, "c++ with typesafe tagged unions" is pretty hard to argue against
like, for all of rust's faults, "c++ with typesafe tagged unions" is pretty hard to argue against
December 23, 2025 at 10:14 PM
not even hardcore-functional-land, even just basics like enums and list processing
like, for all of rust's faults, "c++ with typesafe tagged unions" is pretty hard to argue against
like, for all of rust's faults, "c++ with typesafe tagged unions" is pretty hard to argue against
i mean i suck at all the konami games so we’re even i suppose
December 22, 2025 at 5:09 PM
i mean i suck at all the konami games so we’re even i suppose
but this got me thinking about a language that didn't have closures but could still model functional-style data transformation pipelines by inlining the map/filter/etc chain together with the function pointers' contents...
December 11, 2025 at 10:04 PM
but this got me thinking about a language that didn't have closures but could still model functional-style data transformation pipelines by inlining the map/filter/etc chain together with the function pointers' contents...