Webassembly articles

How We Built Anagram Architect: Meaningful Anagrams with Rust and WebAssembly

An exact anagram uses every letter from the original text once and only once. That rule sounds simple. The difficult part is finding a result that sounds like something a person might actually say. Consider a few of the phrases we used while developing Anagram Architect: Osama Bin Laden → Old man in a base The meaning of life → The fine game of nil Acorn computers → Crap on customer President …

Does Your Word Counter Upload Your Text? A Guide to Private Text Analysis

A word counter looks like one of the simplest tools on the web. You paste in a draft, read a number, and leave. But the text inside that box may be anything but simple. It could be an unpublished article, a client proposal, a job application, a legal draft, meeting notes, or the first version of a book. Before pasting that material into any website, there is a useful question to ask: Does this …

Why an Absurdle Solver Is Technically Difficult to Build

Absurdle looks almost exactly like Wordle: five-letter guesses, gray/yellow/green tiles, and eventually one completed word. The adversarial part is invisible. Behind the board, Absurdle begins with a pool of possible answers rather than committing to one secret word. After each guess, it returns feedback that preserves a difficult compatible family. Every reply remains consistent with the full …

A Mortgage Calculator Built for Decisions: Five Scenarios, Zero Tracking

A mortgage calculator can give you a correct monthly payment and still leave you with the wrong impression. The payment is only one part of the decision. A lower payment may come with a longer term and substantially more interest. A higher payment may shorten the loan. A small recurring extra payment may change the payoff date. A bi-weekly schedule may look attractive, but only if you understand …

How We Built a Four-Board Quordle Solver with Rust, WebAssembly, and Entropy

Quordle looks like four Wordle boards placed beside one another. That description is visually accurate—and computationally incomplete. Every guess is shared across four independent answers. Each board returns its own green, yellow, and gray pattern. A useful next move must therefore do more than fit one set of clues: it should reduce uncertainty across every board that remains unsolved. That is …