Program synthesis is just magical. The next version of https://github.com/Wilfred/suggest.el will support nested functions. E.g. searching 0 => 3 here.
miniblog.
Rust is going to make ! (which represents a function never returning) into a fully fledged type! https://github.com/rust-lang/rfcs/blob/master/text/1216-bang-type.md
Fun post showing how to emulate double precision with single precision: https://web.archive.org/web/20110831040222/https://www.thasler.org/blog/?p=93 (illustrated with mandelbrot fractals)
The IKEA Effect on Software
https://medium.com/@khasan222/the-ikea-effect-on-software-95240dd58a3
(interesting perspective: "I've just built this so you should use it!")
Typescript 2.4 is out! https://blogs.msdn.microsoft.com/typescript/2017/06/27/announcing-typescript-2-4/
The nice thing about type checkers as libraries is that they can be improved over time.
I wish I learnt about cl-defstruct earlier. A little struct with a few keys is really useful for complex return values in (e)lisp.
I've created a homepage for the London Emacs meetup! https://london-el.github.io/
Naturally, it includes an org-mode/zenburn style in pure CSS.
Understanding ResourceT
https://www.fpcomplete.com/blog/2017/06/understanding-resourcet
(covers how it works and the surprising number of pitfalls)
Easy Parsing with Parser Combinators
https://www.lihaoyi.com/post/EasyParsingwithParserCombinators.html
Lovely introduction to parser combinators with a good rationale discussion.
Doing a brute force search of bitwise operators to produce more efficient code: https://www.oilshell.org/blog/2016/12/27.html
Interesting post on GC implementation:
https://olshansky.me//gc/runtime/dlang/2017/06/14/inside-d-gc.html
(I learnt jemalloc doesn't only offer reserving powers of 2 — avoids wastage)
*Amazing* blog post on modelling compilers as a search problem on a Value State Dependence Graph: https://jamey.thesharps.us/2017/06/search-based-compiler-code-generation.html
Making smallest possible test suite using Z3
https://yurichev.com/blog/set_cover/
(really elegant application of the set cover problem to test suites)
Is there still value in learning/doing research in efficient interpreter implementation? Sooner or later you will need bytecode/JIT/AOT.
I'd assumed a lisp-2 required funcall, but MacLisp provided a different solution. ((if t '+ '*) 1) was legal! Feels much closer to a lisp-1.
Showing 3,481-3,495 of 7,508 posts