New OS rule. No threads. Onky lightweight processes and cheap IPC.
Memory wasn’t meant to be shared.
4
Replies
0
Boosts
Excellent!
tbh, IMHO I think "time sharing" was also an anti-pattern that those in our past explored due to budget and limited resources and attempting to maximize what they had available.
I sincerely wish more individuals got to use things such as Amiga Workbench when it was still being actively developed. Having pre-emptive multi-tasking, without user abstractions? Was glorious! Users exist outside of computing; the entire realm of "privilege escalation from standard user to root/admin/etc." is a class of vulnerabilities which fundamentally: do not exist on OSes like Amiga Workbench.
Yet, we'll keep dragging the dead horse of UNIX and Linux and timesharing multi-user concepts into "the future" apparently?
The UNIX Hater's Handbook, as far as I understand, was largely written by Lisp Machine users. Maybe now that https://www.symbolics.biz/ is supposedly open sourcing Genera for non commercial hobbyists, we'll see some of its yesteryears' wisdoms elsewhere as well? Despite LaurieWired's recent YouTube video on it, their name of a "Lisp Machine" was a misnomer, and they ran more than Lisp; there were C compilers (plural, e.g. Symbolics C, Zeta-C), some of which apparently flagged and alerted on out of bound memory access attempts and other programming languages available for Lisp machines as well. Supposedly many of X Windows' early bugs were identified and remediated thanks to Lisp Machine users. The hardware? Was prohibitively expensive. Heck, the last time I saw a CD-ROM of (Open)Genera on eBay it sold for at least $5000 and IIRC required a DEC Alpha to run (though others created a shim layer to emulate such things on Linux systems on AMD64; it seems as if contemporary Genera can now run on AArch64 and more).
Xous' paradigm of memory allocation and access seems fascinating too, though so far I have mostly only read about it; not attempted to run it.
@teajaygrey @petrillic
The Unix paradigms may be quite technologically dated, though they are still better than the vast majority of commercial alternatives, past and present. But that is entirely orthogonal to whether multiple users are supported. The Amiga operating system could have easily been written to support multiple users, without compromising its usefulness or user interface in any way.
For all the numerous faults of Windows, adding multi-user support in NT didn't hurt anything.
I started to reply and wrote way too much.
I definitely appreciate UNIX variants' CLI job control relative to Windows' or (Open)VMS or MPE as some other examples which come to mind.
Workbench as you suggested, could have been extended for multiple users too; though I am still uncertain if such things wouldn't be better served by multiple partitions, or in maybe VMs in todays' more hypervisor savvy realms.
I know I am still dissatisfied with computing.
I am less certain if I have the energy or resources to meaningfully improve much of anythiing for myself or others; though that has been my general intention over the decades, increasingly it feels as if I am more expert than ever, and more than ever, it's overwhelming in all the wrong ways. Like a never ending uphill battle, whereas decades ago, things were smaller, simpler, less dystopian; more fun.
Alan Watts used to talk about "the candle being worth the flame" and that seems as if it is an optimistic sentiment, yet I feel as if I end up more often than not in realms of: "burning the candle at both ends and in the middle too and ouch and why is this happening?"
@teajaygrey
I don't think Workbench would have needed any changes. It would still have been a fundamentally single-user UI. If there had been a multi-headed Amiga system for multiple users, each user would have run their own instance of Workbench.
