On building business models on top of open source projects: https://medium.com/@johnmark/open-source-business-models-considered-harmful-2e697256b1e3
(Different things have been tried historically, but businesses need to provide value! FOSS is not necessarily a business opportunity.)
miniblog.
Allowing users to redefine anything at any point has far reaching design consequences, and has helped lisp/Smalltalk IDEs: https://news.ycombinator.com/item?id=16586347
"for the most part, our members are watching two-and-a-half episodes at a time"
"Our members tend to start with one show and finish it."
Data driven decisions for Netflix when producing content!
https://www.rollingstone.com/tv/tv-features/why-netflix-dramas-sag-midseason-cindy-holland-interview-707986/
TIL Mosaic was not the first web browser, but it played a pivotal role in popularising the world wide web and making it accessible to normal computer users.
Programming languages differ massively in the range of numeric types they offer.
For example, Rust has an extensive selection, targeting machine word sizes and FP hardware (isize, i32, f64) plus unsigned variants.
On New Year's Eve, my family prefers making predictions about the coming year rather than resolutions.
We found this wonderful article of people 100 years ago also making predictions about today, some surprisingly accurate! https://futurism.com/2018-looking-back-century
Insightful article on the economics of Netflix's content business model, and how users consume it differently to historical content providers: https://redef.com/original/netflix-isnt-being-reckless-its-just-playing-a-game-no-one-else-dares-netflix-misunderstandings-pt-3
Refusing to implement impossible features: https://blogs.msdn.microsoft.com/oldnewthing/20050607-00/?p=35413
Selling electronic goods with instructions used to be the norm. It's much rarer now.
What changed? Did consumers stop reading them? Did the average consumer ever read them? Why is it acceptable to skip a manual now?
Chess is beaten (machines can play a perfect game) when there are up to 7 pieces on the board.
This might seem like a simple question of compute power, but clever implementation matters. If you can effectively compress your DB, you can use a bigger one!
https://en.wikipedia.org/wiki/Endgame_tablebase#Computer_chess
Board games seem to have a tendency of giving the first player an advantage, at least with perfect play: https://en.wikipedia.org/wiki/Solved_game
Looking at the increasing dominance of FOSS software in chess programs, I wonder if open source, commoditised implementations are the likely end goal for most software applications.
Perhaps professional programming will become largely glue/business logic?
Optimising the Dart VM, performance cliffs, and the amazing number of string representations in V8: https://mrale.ph/blog/2016/11/23/making-less-dart-faster.html
Contributing etiquette for open source, or why patches should honour the original coding style (however ridiculous): https://tirania.org/blog/archive/2010/Dec-31.html
One crucial factor in Stockfish becoming a leading chess engine is its test framework.
Summary: https://en.wikipedia.org/wiki/Stockfish_(chess)#Fishtest
Initial announcement: https://www.talkchess.com/forum3/viewtopic.php?start=0&t=47885&topic_view=flat&sid=b5662f579ad4c7bea91668bb9d9723a4
Lots of engines have clever programmers, but a strong test technique gave it an extra edge!
A reddit bot that applies machine learning to chess game pictures and extracts the game position! https://github.com/Elucidation/tensorflow_chessbot#reddit-bot
Neat concept. You could extend it to analyse the position too.
Advocacy is important for new tech. People are invested in what they know and use. For them to change, you need to be *persuasive*.
We are demanding ever more tasks from our frontend build processes:
https://css-tricks.com/annotated-build-processes/
(No wonder they can be complex!)
The read function (lex+parse) in lisp is really handy for building little analysis or metaprogramming tools. The runtime has it, so why not allow users to use it?
However, there's no equivalent for just lexing AFAIK. It's a shame because the same arguments apply!
Promoting NO_COLOR as an informal convention for controlling CLI output: https://no-color.org/
Showing 2,321-2,340 of 7,508 posts