Post

austin 🇵🇸
@stdlib.bsky.social

Windows Subsystem for Linux v1, based on the idea of implementing Linux syscalls on top of ntoskrnl.exe, was cool as an idea. It also turned out to be awful in practice and barely worked for anything slightly off the beaten path. I have no idea why people eulogize it so fondly

0 reposts11 likes

austin 🇵🇸
@stdlib.bsky.social

I actually do respect the attempt, not because it's "cool" (it is a cool idea) but because I appreciate studying the negative result. Sometimes you make a bet, and sometimes it is a bad bet and you lose. That's valuable and you don't need to put any mustard on it or anything to see why.

austin 🇵🇸
@stdlib.bsky.social

Most of the impressions I get though when people bring it up are just sort of like, that was cool and this (run a vm) is boring. Perhaps, but figuring out _why_ it didn't work is also cool, if you ask me. There's nothing really lost, only some things to learn.

Catherine
@whitequark.org

this is why i really liked it, and why i paid so much attention to the negative result i think solaris hit something similar?

Catherine
@whitequark.org

there was at least one (i suspect more) talks about how immensely painful the attempt to make them compatible is

the tensor gnomes of false Berlin
@vivsha.ws

i liked it because it was another step in Linux's slow march to world domination. year of Linux on the desktop!! ...but i wouldn't say i eulogize it, or think 1 was better than 2. all i care about is that every dev can get their hands on a nice standard *nix shell and toolchain, regardless of OS