One weakness of online discussion is protracted debates or arguments. What if commenting systems only allowed one comment per user?
(Assuming a robust mechanism to prevent duplicate accounts.)
miniblog.
'Golden tests' are tests that deliberately save expected output to a file, to make them easy to examine or regenerate.
https://ro-che.info/articles/2017-12-04-golden-tests
I'm familiar with the concept, but it's nice to see a handy label for it.
If you're interested in alternative approaches to computing, I strongly recommend exploring the ideas in the Tunes project.
There's a good overview here: https://transistorski.wordpress.com/2010/06/26/better-computing-pipe-dreams-with-tunes/
GitHub is exploring some really interesting editor designs with Xray: https://github.com/atom/xray/blob/master/README.md
It uses CRDT to represent text (rather than rope data structures) and mixes web tech (primarily JS) with Rust to produce a design that's fast but still extensible.
Whilst I've seen many teams use monorepos successfully, it assumes that there's a common contribution process for everyone.
A PL community needs to have different projects with different processes, and that requires packaging tooling. I wonder what the size upper bound is.
Thoughtful criticisms of git from the sqlite developers: https://sqlite.org/whynotgit.html
Great introduction to the AV1 video codec, inter/intra frame compression, and exploiting correlations to save bytes: https://people.xiph.org/~xiphmont/demo/av1/demo1.shtml
The Xiph posts are always good with great explanatory diagrams/sample pictures.
This is very cool: dynamically adjusting an Emacs theme based on the amount of ambient light: https://matthewbilyeu.com/blog/2018-04-09/setting-emacs-theme-based-on-ambient-light
Wow, gcc's codebase is amazingly lispy in places: prefix syntax with parentheses, foo_p for boolean functions.
https://github.com/gcc-mirror/gcc/blob/master/gcc/match.pd
It's natural to worry about the practicalities of new designs. There's a wonderful discussion in https://medium.learningbyshipping.com/nikon-versus-canon-a-story-of-technology-change-45777098038c about camera owners worrying about *batteries* being unreliable. Times have changed!
Good article on RSS on Techcrunch: without helping users with curation, or giving publishers metrics on what is read, it will continue to be supplanted by competitors: https://techcrunch.com/2018/04/07/rss-is-undead/
Git 2.17 has some really nice diff tweaks: colouring lines that have moved without modification, and even showing context by function rather than just N lines! https://blog.github.com/2018-04-05-git-217-released/
Fun post on building IDE features into Excel: https://blog.querystorm.com/index.php/2018/04/04/whynow/
Paxedit is a nifty Emacs project that goes beyond paredit: https://github.com/promethial/paxedit/blob/master/readme.org
It understands e.g. (setf x 1 y 2), where you want to transpose pairs of expressions. It's also smart enough to allow you to operate on the sexp containing point, without you needing to move!
The death of Dennard scaling, Moore's law, and the future of ISA design and performance: https://www.eejournal.com/article/fifty-or-sixty-years-of-processor-developmentfor-this/
(I loved Patterson book and he's the definitive authority here!)
Today I'm playing with Pollen, a 'programmable publishing system'. It lets you intermingle prose and code definitions. https://docs.racket-lang.org/pollen/Backstory.html
Racket promotes defining new languages, so it's a great fit there. The syntax feels lighter than TeX too.
Seen a GitHub project that you want to send a quick patch to? I'm really digging github-clone.el, to quickly clone and fork a project without leaving the comfort of Emacs! https://github.com/dgtized/github-clone.el
Until 2015, LLVM just unconditionally put the function prologue at the beginning of each function, even when it wasn't needed! A friendly introduction to the shrink-wrap pass here: https://reviews.llvm.org/D9210
Every browser has its own distinct URI schema for the "Reader View"! https://www.ctrl.blog/entry/browser-reading-view-uri
I love the idea of mutation testing, but changing my code in arbitrary ways, then running it, is scary. Many PLs don't have a good sandbox available.
Showing 2,821-2,840 of 7,508 posts