From 088a18410e43a209a4864bccad5ddb89a7ee1d5d Mon Sep 17 00:00:00 2001 From: delta Date: Wed, 17 Sep 2025 23:29:44 +0200 Subject: asides, hacker webring stuff and some more --- blog/a_cautionary_tale.md | 3 +++ 1 file changed, 3 insertions(+) (limited to 'blog') diff --git a/blog/a_cautionary_tale.md b/blog/a_cautionary_tale.md index 8a74c00..b9797bf 100644 --- a/blog/a_cautionary_tale.md +++ b/blog/a_cautionary_tale.md @@ -10,3 +10,6 @@ Thus, with my cat ears and thigh highs on, off I went in search of resources on We spent about an hour discussing a solution to my predicament, and in the end we came to the conclusion that while a cubic easing function could be approximated by tweaking the intro and outro durations and altering the easing function, I should simply use another library that doesn't use derivatives for specifying the easing function. This stemmed from the fact that for every well-defined cubic Bézier curve (i.e., `A = (0,0)`, `D = (1,1)`, and `B` and `C` are both in the range `[0–1]`), both the start and the end of the function would be at `y = 0`, making it incompatible with rubato's model (except when intro = 1, which would defeat the purpose of using rubato as an interrupt-friendly interpolator in the first place). The moral of the story? Don’t rice, kids. + -- cgit v1.2.3