the amount of interesting DSLs coming from the #rstats community is genuinely amazing. TIL about @djnavarro ‘s jaysire package for behavioral experiment design https://djnavarro.github.io/jaysire/
Prikaži ovu nit
-
You can see how it could be combined by a description of the statistical design, and you’d have a DSL that end-to-end describes, runs, *and* analyses an experiment. It is mind blowing to think we’re close to being able to do this
Prikaži ovu nit
What should be humbling about this for CS educators is that in CS classrooms, DSLs are still thought to be esoteric and fancy. But to paraphrase Ivan Sutherland, the #rstats community as a whole doesn’t know this is supposed to be A Hard Thing, so they just do it! Incredible.
-
-
In my view, the
#rstats community works really hard to make everyone understand what’s going on, and how uniquely empowering this isPrikaži ovu nit -
Like, I honestly look at this and it’s the closest community I know of to papert’s vision. I know this is gushing, but I felt this last year and I feel it again this year. It’s so damn refreshing
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
-
-
Like, let me give you another example I also learned of today. There's this package in R called tfautograph which takes an R function, analyses it, and then emits tensorflow code that "does the right thing":https://t-kalinowski.github.io/tfautograph/articles/tfautograph.html …
Prikaži ovu nit -
in CS-speak, this is a "source-to-source compiler"; in R, this is an (admittedly awesome) package that someone built to solve their problem!
Prikaži ovu nit - Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.