TL;DR: In a recent camping trip we noticed just how cold our propane tank was getting after cooking for a while - we started out pretty chilly anyway (about 50 F). The physics of this is cool and I wanted to play with modeling it.
I have one of these little two burner propane stoves that we use when camping, it looks something like this:
While we were cooking breakfast and boiling water for coffee, we noticed that the propane tank was getting super cold - there was not only a thick layer of frost forming on the outside, but a small chunk of solid ice on the bottom of the tank where it met the picnic table…
Here are some podcasts I’ve been into lately and why I like them.
Note - I’m not affiliated with any of these podcasts, I’m just a listener.
Tech Backend Banter - Lane from boot.dev interviews a variety of other devs (mostly) and talks about their experiences and projects. I particularly like that it’s not only Golang focused. It’s mostly entertainment, but I’ve learned a few things from it. I really enjoyed the episodes with John Crickett about ‘mechanical sympathy’ (and also his coding challenges) and the one with Thomas Ballinger about convex.
TLDR - My general takeaways from working on two similar apps. One the OG server-returns-html way with FastAPI/HTMX and one with Next.js using the app router / RSC / SSR. These are my quick take aways from the point of view of working on them from the perspective of a “non-web-dev dev”.
Intro I am currently working on two somewhat similar web apps, one using FastAPI with HTMX and one using Next.
TL;DR: A quick “non-mathematical” introduction to the most basic forms of gradient descent and Newton-Raphson methods to solve optimization problems involving functions of more than one variable. We also look at the Lagrange Multiplier method to solve optimization problems subject to constraints (and what the resulting system of nonlinear equations looks like, eg what we could apply Newton-Raphson to, etc).
Introduction Optimization problems are everywhere in engineering and science. If you can model your problem in a way that can write down some function that should be minimized or maximized (the objective function) to get the solution you want, even in cases where there is no analytical solution (most real cases), you can often obtain a numerical solution.
TL;DR: The S63 encryption scheme is old and can be broken pretty easily depending on the information you have. These are some notes on how with a simple Python script to show what I mean.
So this is an old standard that will soon be replaced by S100, but it’s still in place and used today. The encryption part of the scheme is mostly to allow data suppliers to license their data to users in a standardized way and collected payment for access to that data.