@mcc

I was seriously considering installing ReactOS a month or so ago but when I looked into it it appears ReactOS has just never got that Big Community Love and the compatibility of ReactOS is worse than modern Wine.

yeah that doesn't surprise me, ReactOS has always been like that. i think the idea of “why not just use Linux underneath” is a long-floated idea in that community actually, i don't know if anyone actually did it though

when MS pulls the plug on Win10 there will be demand for that

you understand my motivations perfectly :)

Maybe u could even reuse the ReactOS shell. Seems "obvious" the ReactOS GUI probably runs in Wine

I WAS ALREADY PLANNING TO DO THIS WE SEE EYE-TO-EYE

@mcc it seems so obvious and so doable.

i do have a slightly grander additional goal though: i also want to figure out how to make this os have decent security. an os that is designed to let you download and run random exe's from the internet is great in many ways but not in security, at least not by default. but i have some ideas. i would want the user to be in full control of this though!

@hikari it seems like there are lots of sandboxing options that have never been seriously explored in industry because the people who control the sandboxes on consumer devices have a greater interest in limiting what the users do than in allowing the users to do a broader range of things safely

@mcc
i think react's main problem is that the kernel is pretty primitive compared to linux
support for smp isn't something you should try and add after decades of work
@hikari

@7331 @mcc oof i didn't know they hadn't got that far, but this isn't so surprising really. it's a hobby project for windows kernel enthusiasts with no prospects of enterprise support (they pay for windows), and kernel enthusiasts are a tiny group that probably mostly work for microsoft, work for driver vendors, have been contaminated by research access to microsoft code, or… work on malware

@hikari
as a kernel enthusiast (TM) i can tell you that yea, there's not too many people who like to work on that sort of stuff (and most - like me - just roll their own because it's way more fun and even more abandon their projects early on because it's a lot of work) and i can't imagine it was a lot easier in '96 xD

so i guess i can see why features that did have an immediate impact on usability (e.g. writing a usb stack) were prioritized over getting the architecture right but the problem there is that it's extremely hard to go and change this sort of stuff later on (a lot harder than changing the architecture of normal software imo) and even harder if you want your implementation to be performant and sane
@mcc

@7331 @mcc yeah. i also think people don't appreciate enough that big architectural changes are fucking hard work and the most reliable way to get them to happen is to pay a team of people to work full-time on it. it's very difficult for actual hobbyists to get that stuff done