4 ms·
You have it backwards from where I'm standing. 'null' (and to a large extent mutability) drives a gigantic hole through whatever you're trying to prove with co
by mrkeen 1y ago
You have it backwards from where I'm standing.
'null' (and to a large extent mutability) drives a gigantic hole through whatever you're trying to prove with correct-by-construction.
You can sometimes annotate against mutability in OO, but even then you're probably not going to get given any persistent collections to work with.
The OO literature itself recommends against using constructors like that, opting for static factory pattern instead.
- munificent 1y agoNullability doesn't have anything to do with object-oriented programming.
- mrkeen 1y agoYes, yes, "No true OO language ..." and all that. But I'm going to keep conflating the two until they release an OO language without nulls.