2022-10-22 10:17:22 I once was working on a "semantic" web browser that would show all websites the same way, disregarding CSS/JS. Never finished it. But maybe Forth could be a neat middle-ground for web crawling and getting latest news from a handful of sites. 2022-10-22 11:33:49 Btw, what's Forth's relation to SIMD chips? 2022-10-22 11:34:49 what's Ada's relation to FPU chips? 2022-10-22 11:38:04 Ada doesn't have floating point? 2022-10-22 11:39:05 it does 2022-10-22 11:39:18 theyre completely unrelated 2022-10-22 11:39:19 much as forth to simd chips 2022-10-22 11:43:11 it could be fun to do a 3-stack forth, where you've got an extra vector stack, but i dunno if that would actually be, like, ... good 2022-10-22 13:05:54 surely you'd just drop the addres of your table of vectors on the stack and have at it 2022-10-22 13:24:00 On google there's one forth clone doing simd on each instruction :d 2022-10-22 13:25:48 Some shader thing 2022-10-22 13:26:01 https://github.com/janm31415/vectorforth 2022-10-22 13:26:24 "if you compile 1 2 +, then the result will obviously be 3, but actually you will get 16 times the result 3 (using AVX-512), " 2022-10-22 13:31:00 that seems wasteful 2022-10-22 13:31:35 Electricity? 2022-10-22 13:31:38 Time is the same 2022-10-22 14:39:04 Ive wondered if you could process multiple stack items at once with that to get a speed up 2022-10-22 14:39:21 Unless they depend on each other, I guess 2022-10-22 14:39:28 1 2 3 + + won't work 2022-10-22 14:39:47 but maybe 1 2 + 3 4 + + would work 2022-10-22 14:39:56 Yes :) 2022-10-22 14:40:10 You "just" need a notation to tell the compiler about that relationship... 2022-10-22 14:40:57 ya or do some static analysis 2022-10-22 14:41:10 which is not very popular in forth for anything :) 2022-10-22 14:43:51 "the electronics fried itself after we applied static" 2022-10-22 14:44:03 could see why that isn't popular