Why does Haskell say Nothing instead of None? Didn't occur to them or was rejected? I can see why Some is less than ideal since it rhymes with "sum".
Conversation
And I presume Rust far postdates these Haskell decisions. But I still suspect lots of options got discussed.
1
1
Rust borrowed the naming of Option, Some, and None from OCaml and SML. Not sure where Maybe, Just, and Nothing came from - maybe Miranda?
2
2
Hmm, I can't see maybe defined in Miranda's stdenv (cs.kent.ac.uk/people/staff/d), but it might have been a datatype commonly in use? That said, idiomatic Miranda and early Haskell leaned much more on partial functions, so it might have seen less use.
1
2
Another Haskell precursor, Orwell, does not define it in its stdenv either: homepages.inf.ed.ac.uk/wadler/papers/
1
2
You might need to go through the early Haskell papers and try to spot where it first appears 🤔
1
2
Thanks much for the digging you've done.
1
2
1
2
Interesting - I don't think I see a mention of Maybe in v1.0 of the Haskell Report? raw.githubusercontent.com/joyofhaskell/h … unless I'm missing something?
1
2
Going through typeclasses.com/timeline, it pops up in 1995 in web.cecs.pdx.edu/~mpj/pubs/spri - along with an Error type which looks very close to Rust's Result type! (unlike the Either type which was eventually added)
Seems like Maybe and Either were then added to the Haskell report in version 1.3: raw.githubusercontent.com/joyofhaskell/h - so it seems like it is a Haskell thing?
1
2
Would love to know more background on it, but it might require trawling through mailing lists, or emailing Mark P. Jones. Now I am curious about where and when Standard ML got option!
1
2
Show replies