I keep thinking about this.
If "a small set of orthogonal primitives" is good, why does *no* popular language have this?
Possible explanations:
1️⃣ It's not good
2️⃣ It's good, but doesn't drive adoption
3️⃣ Number of primitives should be "just right" (not too big, not too small) twitter.com/dubroy/status/…
If "a small set of orthogonal primitives" is good, why does *no* popular language have this?
Possible explanations:
1️⃣ It's not good
2️⃣ It's good, but doesn't drive adoption
3️⃣ Number of primitives should be "just right" (not too big, not too small) twitter.com/dubroy/status/…
Patrick Dubroy
@dubroy
·
Mar 6
It's often said that programming languages should have "a small set of orthogonal primitives" but I wonder if there's any empirical basis for this?
Another possible explanation:
4️⃣ The process of adoption drives minimal languages to become less minimal.
4️⃣ The process of adoption drives minimal languages to become less minimal.
An objective comparison is tricky, but looking at # of production rules in ANTLR grammars —
Lua: 62
Smalltalk: 81
VB6: 158 + 226 lexical
C: 232
Rust: 240 + 151 lexical
Swift5: 317 + 236 lexical
All from github.com/antlr/grammars…
Lua: 62
Smalltalk: 81
VB6: 158 + 226 lexical
C: 232
Rust: 240 + 151 lexical
Swift5: 317 + 236 lexical
All from github.com/antlr/grammars…
https://twitter.com/dubroy/status/1589562678103535616 ∙ Archived on 2025-03-28.