Casey #Muratori’s keynote about, largely, #disjoint-unions and how it is sad that #C++ doesn’t have them. “I’m not saying OOP was a mistake. I’m saying this [a compile-time hierarchy of encapsulation that matches the domain model] was a mistake.” Also traces the #history of records in #programming-languages through Simula and Hoare’s record-handling paper and the “plex” of Douglass Ross of the MIT Servomechanisms Laboratory in AED, Algol Extended for Design, and also #Sketchpad, which had in-memory records (“chickens”) linked together in circular linked lists, which apparently was an idea he got from Ross (“n-component elements”). Both Ross’s plexes and Sketchpad had function pointers in the records. (But the constraint solver was “the most unencapsulated thing that you could possibly imagine”.) Also documents how Looking Glass in 01998 introduced the #Entity-Component-System pattern in Ultima II Underworld, but really introduced it with Tom Leonard’s Thief: The Dark Project. At 82'40” he points out that he independently invented a worse version of ECS at definitionSIX for Negaman in 01997. “It’s just, I sucked at it, and Looking Glass was good.” Basically his “35-year mistake” thesis is that we almost had ECS in 01963 with Sketchpad, but it took until 01998. He likens 01990s OOP dogma to playing Magic: The Gathering, which I think is insightful. At 136'45” he tells his devotee Ryan Fleury that he has snatched the stone from his hand and may now leave. Nice quote: “You should focus on the hardest stuff. You should say, ‘what solves the hardest problems?’ because we can always then take that and scale it down, and remove things from it or dumb it down, for people to use in cases that aren't as hard. But it’s almost impossible to take something that only solves simple problems and scale it up to something that solves hard ones.”
on 02025-07-22an approach to programming in C using “handles” like filehandles instead of pointers, catching use-after-free errors by dividing the handle into a slot number and a generation counter. Claimed advantages include smaller handles and thus lower memory usage, more reliable detection of leaks and use-after-free, better memory locality, more memory layout flexibility, and no fragmentation. Related to #Entity-Component-System, but each handle only indexes a single component, rather than potentially several as in ECS.
on 02023-06-21