If a PL defaults to immutable lists, we call it 'functional'. But if a PL defaults to immutable strings, we still call it 'imperative'! Odd.
miniblog.
Emacs tip of the day: C-s C-s will repeat your last isearch! Works with anzu too! (via @emacs_knight)
I only test my elisp against 24.4 and later now. Supporting earlier versions is hassle and no-one has asked for it.
Firefox 48 also adds -webkit prefixed CSS properties to the unprefixed ones: https://developer.mozilla.org/en-US/Firefox/Releases/48#CSS (seems silly, but makes sense)
Firefox 48 adds a really neat CSS tool: adjust absolutely positioned elements by dragging them on the page! https://developer.mozilla.org/en-US/docs/Tools/Page_Inspector/How_to/Reposition_elements_in_the_page
Good error reporting in macro systems is tricky. An interesting Rust RFC to provide better messages: https://github.com/sgrif/rfcs/blob/sg-add-error-macro/text/0000-add-error-macro.md
rust-mode has 194 tests: indentation, font-locking, everything. It's really impressive—the Rust community values testing highly (cf cc-mode)
Rich terminal applications: https://ballingt.com/rich-terminal-applications-2 (essentially responsive design for CLI apps! Neat!)
Pervasive docstrings, elisp book, manual and tutorial (all offline and dynamic) make Emacs self-documenting. I miss it on other platforms.
One brilliant thing about macros is that you can write them, then throw them away if they're not worthwhile. I often end up doing this!
Fascinating history of nodeJS's module system: https://github.com/nodejs/node-v0.x-archive/issues/5132#issuecomment-15432598 -- particularly interesting comments on APIs not right the 1st time.
Interesting machine learning application: detecting abusive messages online! https://www.bloomberg.com/view/articles/2016-07-29/i-can-be-baited-with-a-tweet-and-that-s-ok
Coati is a really slick proprietary tool for exploring C++ codebases: https://www.coati.io/#how-it-works
Sometimes when thinking on PL design, I worry about efficiency. Considering the num of successful AST-walking interpreters, it seems silly.
Interesting talk on Fortress: https://www.youtube.com/watch?v=EZD3Scuv02g The parser rejects whitespace that doesn't match precedence!
I still can't make my mind up about using lists for representing syntax. It's great for macros, but ASTs often need more (eg line numbers).
Cuis is the only image-based Smalltalk dialect that supports a PR workflow. https://github.com/Cuis-Smalltalk/Cuis-Smalltalk-Dev I'm hopeful more will exist in future.
Really neat example of using Smalltalk to interactively explore numeric differentiation:
I've just discovered @emacs_gifs by @ocodo -- superb Emacs demos, highly recommended.
TIL about an incredible Microsoft tool that automatically writes code that works with the examples provided: https://microsoft.github.io/prose/documentation/flashfill/usage/
Showing 361-380 of 922 posts