> It has an achingly elegent internal structure, but a user interface that has been asleep for the past decade. [...] I was writing a civilized window manager for it in my spare time
Brutal, but the window manager is probably one of the three big things hurting plan9 adoption (along with the lack of common editors like emacs/vim, and the lack of modern web browsers)
The Plan9/9Front folks must like the look or rio, sam, acme. Otherwise somebody would have written a more modern looking desktop for it.
When Steve Jobs returned to Apple, he brought NextStep with him. The new version of the OS was basically FreeBSD core with a NextStep inspired desktop. The same could have been done for Plan9.
While the look certainly isn't flashy, I'd be fine with the look. It's the reliance on 3-button mice that's a problem for me. At least modern laptop touchpads can do middle mouse click with 3-fingers but that's a pretty lousy UX, so it's really only practical with a real mouse.
Also I'm much more of a keyboard + tiling window manager kind of a person, so the plan9 focus on the mouse isn't my cup of tea.
NeXTSTEP was never a “FreeBSD core”, at least not as is usually implied by that. It was XNU (Mach with FreeBSD’s world bolted on and IOKit shoveled in) with a custom Desktop Environment.
That's not a correct characterization. OS X is entirely NeXTStep with some classic Mac OS UI elements bolted into it. You couldn't just swap some bit out for Plan9.
I would have said the biggest thing hurting plan 9 adoption is that there is no C++ compiler. And this sort of slots into the next biggest thing, it's not posix, Yes posix is several sorts of terrible and a large part of the plan 9 ethos is to break from it's unix past and do things right. But posix keeps everybody on the same page so things tend to be easy to port. It is very hard to port software to plan 9. There is that posix layer APE but it does not appear to help for any moderately complicated program. I think this is because there is no C++ compiler.
Every couple of years I resolve to try plan9 again, because I love the idea of it. and every time I sort of lose interest, I am mot a good enough programmer to make the pieces I want from scratch, it is too hard to port existing software, every thing is too different from any unix system. So I struggle for a day or two then go back to my obsd. As shallow as it is I think having a good web browser would solve 90% of my issues. and we loop back to the no C++ compiler again.
That does indeed sound like what Plan 9 devs would do. Funny how this "no C++ compiler" stuff transferred to Go too, where there is no cppgo, only cgo integration :)
Having said this, yes even though C++ was born on the same build as UNIX and C, there wasn't much love towards it from UNIX authors, nor towards ANSI/ISO C for that matter, hence why Plan 9 and C compiler are their own thing, same on Inferno regarding the provided C compiler.
the hardware support was pretty bad. it ran on x86 with only the most basic of hardware with a vga graphics adapter in some basic vesa mode.
it never felt serious, just more like a technology demo. it was also hard to really see the beauty of the 9pfs on a single host... i always assumed it needed to be on some bell labs network with a bunch of other plan9 stuff to really shine.
I remember trying out Plan 9 a few years ago. I like its underpinnings; I like how the designers pushed the Unix notion of "everything is a file" to its limits, with various services being presented to the operating system as file servers implemented using a protocol named 9P. It's a cool research operating system with a lot of very interesting ideas, but ultimately it's not my daily driver due to a lack of hardware and software support, though I am working on a research project (but for conventional *nix operating systems) that takes advantage of 9P.
Carmack's description of the rio GUI is apt; rio is heavily based on Xerox PARC's Cedar environment from the mid-1980s (https://www.youtube.com/watch?v=z_dt7NG38V4). I had a hard time getting used to rio, though eventually I learned it. I even purchased an old Sun three-button mouse (with no click-wheel) since Plan 9 didn't like my click-wheel mouse. rio does not adopt a lot of the conventions that were introduced by the Apple Lisa, Apple Macintosh, and Microsoft Windows, resulting in a completely foreign experience for most new Plan 9 users. There is a misconception that Apple and Microsoft simply stole the idea of GUIs from Xerox PARC without further innovating on it. This misconception is false; the Apple Lisa and the Apple Macintosh introduced many GUI conventions that were not present in Xerox PARC's GUIs (this video comparing the Apple Lisa to the Xerox Star highlights the differences: https://www.youtube.com/watch?v=pBiWtJJN5zk), and Windows introduced additional innovations. Back to Plan 9, the result is a GUI that has many conventions that were explored by Xerox PARC but were not adopted by the Mac or Windows, such as mouse chording (https://en.wikipedia.org/wiki/Mouse_chording). Hence, there's a learning curve for those accustomed to conventional GUIs.
Even so, the ideas behind rio seem to do a good job with exploiting Plan 9's architecture and showing how it could support a more tools-oriented design. Moreover, it was borrowed from Xerox PARC's Cedar. It would have been quite a research effort to create a Mac- or Windows-like GUI for Plan 9 that made full advantage of Plan 9's "everything is a file" architecture. Merely implementing a Mac- or Windows-like GUI to Plan 9 might not have taken full advantage of Plan 9's architecture, which I think is the key difference between Plan 9 versus other operating systems.
Perhaps in some alternate universe where Apple and AT&T merged in the 1990s and Apple built their successor to the classic Mac OS on top of Plan 9 with Apple people like Larry Tesler, Bruce Tognazzini, and Alan Kay joining forces with Dennis Ritchie, Ken Thompson, and Rob Pike....come to think of it, it's quite interesting to think about that possibility, though I don't think Apple would have made its post-1997 comeback without Steve Jobs, no matter how good "Mac OS Plan 9" would've been.
Brutal, but the window manager is probably one of the three big things hurting plan9 adoption (along with the lack of common editors like emacs/vim, and the lack of modern web browsers)