Twitter | Search | |
Hillel Wayne
Formal methods, software history, chocolatiering. Author of Practical TLA+ (). DMs open, ask me questions!
29,521
Tweets
206
Following
10,238
Followers
Tweets
Hillel Wayne 9h
Replying to @modernserf
Guess what I just lost 3 hours to
Reply Retweet Like
Hillel Wayne 12h
Replying to @modernserf
And now it's 1.0
Reply Retweet Like
Hillel Wayne 12h
So should we just make Factorio game of the year now or keep up a charade of suspense for the next four months
Reply Retweet Like
Hillel Wayne 12h
I posed it to a couple of pure math friends I have and they found it frustrating too. Though I'm also remembering that said some meaningful things here, but I forget what they were exactly.
Reply Retweet Like
Hillel Wayne 12h
Replying to @lexi_lambda @prathyvsh
I'm afraid I don't either. I was mostly inspired by how in a lot of math domains adding negative statements reduces the scope of what you can say. It seems to be a pattern, but I don't know if there's something deeper to it, and if so what it is. My ignorance frustrates me.
Reply Retweet Like
Hillel Wayne retweeted
Alexis King 14h
It’s interesting to me how, whenever I write about capturing invariants in static types, a really common example commenters bring up to argue static types aren’t useful is “a type for a valid email address.”
Reply Retweet Like
Hillel Wayne retweeted
Joe Groff 13h
Replying to @hillelogram
I took a stab at designing a verbose APL a while ago, but didn't get around to actually implementing it
Reply Retweet Like
Hillel Wayne 13h
Replying to @danielrheath
Overwhelming scientific evidence says "go to bed"
Reply Retweet Like
Hillel Wayne 13h
If people aren't going to read comments, why would they read commit messages?
Reply Retweet Like
Hillel Wayne 16h
Replying to @deech
Streams of Text is the Wrong Abstraction for Source Code REPRESENT
Reply Retweet Like
Hillel Wayne 16h
That's a very common assumption! AbDTs seem like a waypoint because they're conceptually simpler, but they're conceptually simpler because they came after. New paradigms usually start out way overcomplicated as people figure out what's essential and what's not
Reply Retweet Like
Hillel Wayne 16h
Replying to @hillelogram
If you enjoyed this, I wrote a bonus section for my newsletter, where I start off talking about permutations and space efficiency and then get distracted by large number approximations. Now featuring LOGARITHMS!
Reply Retweet Like
Hillel Wayne 16h
s/ADT/AbDT
Reply Retweet Like
Hillel Wayne 16h
A better way of putting it (I was sloppy earlier) is that ADTs were an attempt to capture some of the value of inheritance while avoiding some of its pitfalls
Reply Retweet Like
Hillel Wayne 16h
Replying to @stilkov @maxheiber
Yeah, Liskov explicitly cites Simula as part of her inspiration.
Reply Retweet Like
Hillel Wayne 16h
Replying to @maxheiber
Oooh, let me think about this. I know a lot of the rank stuff is really tricky to translate, but to my understanding Julia and Numpy took a lot. But I imagine you're thinking of more fundamental ideas, which I'll have to think about
Reply Retweet Like
Hillel Wayne 16h
Replying to @taralx
Vinegar leg is on the right
Reply Retweet Like
Hillel Wayne 16h
Replying to @michaeljforster
Typo, I meant "inspired by inheritance in early OOP". Keys are right next to each other
Reply Retweet Like
Hillel Wayne 16h
You're not missing anything: Simula was 1967, CLU was 1975, AbDTs went viral shortly after Liskov invented them, so were pretty widespread by 1980
Reply Retweet Like
Hillel Wayne 17h
Replying to @NicolasRinaudo
I feel like this gets more hate from the clean code crowd than the Haskell crowd, tbh
Reply Retweet Like