An efficient user-space IA32 / AMD64 emulator for AArch64. It effectively replicates Apple’s Rosetta 2 on GNU/Linux, enabling you to run IA32 or AMD64 binaries on an AArch64 machine as if they were native executables.
In order to attain performance comparable to Rosetta 2, which Apple manage via special hardware support, Box86 cheats by sacrificing correctness in some places. So don’t run your scientific code on Box86 — games and productivity apps should be better candidates.
A fork of the Checker Framework with support for JSpecify.
The first public release is here. I can hardly wait to add it to my libraries.
A naming convention for CSS-styled HTML classes.
It is short, yet there are so many gems in this one. Recommended.
A book on designing systems using hypermedia.
An alternative explanation.
Of course there are reasons for the other direction, too. They just aren’t in this article.
A slide deck advocating for more Unpoly and less imperative JavaScript.
Extends HTML with an expression language that enables dynamic modifications in a declarative way.
Another htmx-like set of interactive extensions to HTML. Can be used for hypermedia-driven applications.
A hypermedia approach to mobile app development.
For partial updates of your web page, just stream HTML to the client.
Declarative HTML extensions for interactivity that degrade gracefully in the absence of JavaScript. Can be used for hypermedia-driven applications. Compared to htmx it is also older and more comprehensive, but accordingly more complex.
Interactive HTML extensions for hypermedia-driven applications.
A UN*X running on WebAssembly in the web browser.
Medley, the Interlisp development environment, is now free software.
Interlisp represents one of the two branches of Lisp that eventually culminated in Common Lisp, Lisp Machine Lisp/Maclisp being the other. Its development environment was way ahead of its time, including a structured code editor and an a graphical user interface built for interactive development.
Links:
A proposal for a dialect of C that is free of undefined behavior.
Now, undefined behavior is, of course, useful to compiler vendors because it enables certain optimizations. But in my mind it just makes a C a strange language without a clear idea of itself: It is not the portable assembly language that many of its users treat it as, one that you can rely on to translate into predictable hardware-level instructions; yet at the same time it is not a high-level language either, unsafety and machine specifics lurking behind every corner.
The same author later follows up rejecting the idea of achieving broad consensus before implementing such a dialect as unrealistic. Instead, they argue, someone influentual enough should just pave the way by writing their code in a Friendly C and making it the quasi-standard.