4 ms·

This strikes me as revisionist. Research Unix v8-20 had quietly trended towards more things-as-files. The hardware environment that Plan 9 assumed was very diff

This strikes me as revisionist. Research Unix v8-20 had quietly trended towards more things-as-files. The hardware environment that Plan 9 assumed was very different - separate file servers, compute servers and graphic workstations, plus it came out when Microsoft and Windows were at their maximum power. Beyond all that, today's leading OS, Linux, has incorporated many of Plan 9's ideas. /proc filesystem, the signalfd system call, and UTF-8 are all examples. I think it's dangerous to assume what this article advocates. You risk trying to make a faster horse, or maybe the Uber of $X

lproven 4y ago
> Research Unix v8-20 had quietly trended towards more things-as-files Hang on, what? AFAIK UNIX 10 was the last ever version. https://en.wikipedia.org/wiki/Research_Unix#Versions https://en.wikipedia.org/wiki/Research_Unix#Versions
zokier 4y ago
> Research Unix v8-20 had quietly trended towards more things-as-files That brings up another point; UNIXes v8-v10 are even more irrelevant and forgotten than plan9, despite presumably being more compatible with the venerable v7 UNIX. So pinning plan9s failures on it being different seems to me very misguided
zozbot234 4y ago
Linux is still incorporating ideas from Plan 9 in a haphazard and unplanned way. "Containerization" (kernel-level namespaces, really) was a big one; user-space block devices was one of the latest. I do think that Plan9's de facto requirement for graphics-capable hardware is a bit weird in retrospect. It really was made to be a research OS for networked workstations (and the default UI basically clones Oberon) even though the design concepts it ultimately leverages are far more applicable.
- pjmlp 4y ago
  IBM mainframes got there first.
- MisterTea 4y ago
  > I do think that Plan9's de facto requirement for graphics-capable hardware is a bit weird in retrospect. Graphical workstations were always the desired end result for decades since the mother of all demos. In the 90's GUI's were all the rage: Windows 95 removed the need for DOS and MacOS was already leading in this space. So it makes sense that plan 9 was designed to be graphics first.

lproven 4y ago

> Research Unix v8-20 had quietly trended towards more things-as-files Hang on, what? AFAIK UNIX 10 was the last ever version. https://en.wikipedia.org/wiki/Research_Unix#Versions https://en.wikipedia.org/wiki/Research_Unix#Versions

zokier 4y ago

> Research Unix v8-20 had quietly trended towards more things-as-files That brings up another point; UNIXes v8-v10 are even more irrelevant and forgotten than plan9, despite presumably being more compatible with the venerable v7 UNIX. So pinning plan9s failures on it being different seems to me very misguided

zozbot234 4y ago

Linux is still incorporating ideas from Plan 9 in a haphazard and unplanned way. "Containerization" (kernel-level namespaces, really) was a big one; user-space block devices was one of the latest. I do think that Plan9's de facto requirement for graphics-capable hardware is a bit weird in retrospect. It really was made to be a research OS for networked workstations (and the default UI basically clones Oberon) even though the design concepts it ultimately leverages are far more applicable.

pjmlp 4y ago
IBM mainframes got there first.
MisterTea 4y ago
> I do think that Plan9's de facto requirement for graphics-capable hardware is a bit weird in retrospect. Graphical workstations were always the desired end result for decades since the mother of all demos. In the 90's GUI's were all the rage: Windows 95 removed the need for DOS and MacOS was already leading in this space. So it makes sense that plan 9 was designed to be graphics first.