Side note, why were Ticketmaster hiring systems programmers writing such low level code? I grew up in the 90s/00s so missed this era of computing.
At some point in the 1970s or 1980s, Ticketmaster made the extremely unusual decision of writing their own operating system for VAX. Even back then, that was a pretty weird decision. And it wasn't Terry's idea, they started doing it before he worked there.
https://news.ycombinator.com/item?id=9808995
And apparently decades later their backend was still stuck on it (is it still stuck on it???), albeit now running under a software emulator instead of physical VAX hardware.
Most people on VAX ran either Unix or VMS, both of which offered migration paths to other platforms (Alpha -> Itanium -> x86 for VMS; and there were lots of alternative hardware platforms for Unix). But Ticketmaster's weird decision to write their own OS made that almost impossible.
Also, the parent commenter's hypothetical does not take into account that the "moderately competent systems programmer" would have to do this work while suffering from a debilitating mental illness.
To be clear, this is not simply a question of "Given 10+ years with no other committments, could someone else do this work?" The question is whether another person with schizophrenia has actually done this work, staying focused for such an extended period.
(If it's common for employed "developers" to have schizophrenia, then please disregard this comment.)
Not discounting his achievement though, it is still noteworthy.
It deserves to be talked about. The cult following and media attention that he got is really just a function of the romanticizing of mental illness and the "crazy genius" trope that people projected on him.
Writing a decent operating system is a good technical achievement but quite a few people have done so as hobby projects. The real attention came from this "temple for god" thing. Instead people should have talked about what it was, a mentally ill person that had no help for decades and it cost him his life. It's quite astonishing and telling that with all the attention he had, he ran around homeless and confused while people were treating him like a circus curiosity or fringe religious character.
But could they do it while suffering from severe psychosis and believing that they were being hunted by bioluminescent CIA operatives?
What I like about TempleOS is that Terry wasn't afraid to make something new and unique, even though it was completely incompatible with everything else.
DolDoc is such a cool idea. Not a unique one, but you don't find many operating systems which have a shell based on that. Since the command output is in DolDoc, you can type `DIR` at the command line and get a list which expands into the whole file tree when you click on it. DolDoc also means that his games have all the images and models embedded in the code, and you can just go there and edit them.
And there are a lot of these. Sure it's clunky and weird and sort of unusable, but I appreciate it for what it is. And I think it's really awesome. Take a look at a random person's osdev project and you'll see mostly the same thing — a unix-like that tries to be compatible with POSIX and port all the utilities. TempleOS is as far from that as you can get (though definitely not unique either).
(Also, a minor correction: TempleOS is not real mode, it runs in full 64-bit mode with symmetric multiprocessing and everything. The text UI isn't the VGA text mode, it's a framebuffer that also displays graphics).
I think this also touches on the other thing that makes TempleOS cool. It was designed as a modern C64, and a lot of the weird design decisions come directly from that. It's a hacker's operating system and it shows.
I seem to remember seeing a different take on hackable operating systems, something written in Scheme with a PreScheme core I think? I remember thinking it was pretty cool. The dynamic nature of Scheme allowed you to have access to pretty much anything you want, but also run programs with less privileges.
Of course there's also DuskOS. It's written in Forth which I like, but I'm not the biggest fan of their dialect. I'd like to attempt something in a more classical Forth, but I never get to it. One of these days...
A bit like picassos answer to someone saying to him about his art "I could have done that!!" being "why haven't you?"
I would put templeos at the same level as art in that it's the creation and life rather than the craft or utility. Or in other words there is something transcendental about the project which would go beyond (but not technically) something a competent systems programmer should write as a hobby.
Art Vs craft has more of an artist inside of it. Noone else could have written templeos.
EDIT: to be clear, that doesn't mean it's not impressive or not a great achievement. It's a fantastic piece of art IMO.
It opened him up to work on building/promoting GNU and FSF.
I have said for years that things like Free software would have come along eventually, he just happened to be in the right position to make it happen when it did.
I never worked with Solaris and the like, but it's my understanding the GNU tools were useful enough that they got routinely installed on the commercial unixes before Linux came along.
This is sadly something that seems to be getting lost. GNU tooling doesn't seem that competitive anymore, and the trend is to write newer, fancier replacements, like ripgrep for example.
I am still 100% on board with the libre/freedom stuff but you are right, a lot of the GNU stuff is being over run by more nimble newer things. A part of that is probably in part due to the technical legacy of some of the GNU tools having code dating back to the 1970's.
You can find such projects in sourceforge.
What practical value does it bring outside of 15 minute awe from internet nerds? No sarcasm.
I think that's exactly the point - it doesn't. That's why it's an artpiece.
That’s how art works. And computers are confined to Solving Business Problems and Delivering Value To Shareholders. What a boring existence that’d be. On an unrelated note, I don’t live in Silicon Valley.
I met Terry online in a couple OSDev forums in the early 00s, before his downward spiral. He was already quite combative, but still coherent. This was even before he named it LoseThos (he changed the name a few times after). I used and gave feedback.
Anyway: There were about a hundred people back then writing hobby operating systems. Perhaps everyone in those forums was? I was one of them. Some OSs had graphics. Some had POSIX. A few people could even run Firefox. I never got to this point. Some even had that same C64-like character of TempleOS.
Like GP, this is not to trivialize the work and dedication that went into TempleOS. But unfortunately he didn't became famous because of it at all. Most reddit people who talk about TempleOS haven't even installed the OS, and pretty much everyone that mentioned TempleOS to me in real life haven't even seen Youtube videos of it.
Which is very sad, because IMO lack of people using it is why Terry ended up becoming isolated and combative in those communities. People still care more about a cool story than a labour of love.
Very different to Competent Systems Programmer writing Unix Clone #983
It had a compiler for the script, a bytecode interpreter and VM, and different HAL's depending on the system in use, with the VM being used to play what were then called "multimedia" titles - graphics and text and sound, all on the same screen. This all had to be done by me, a lowly junior, and when it was finally delivered and titles were shipped, I moved on to other things .. an online version of the engine that could get its bytecode over ethernet .. and I had to write all the drivers and HAL code myself.
Sure, this was the 80's and programmers were a different breed then, but it always bugs me when Mr Special 21st Century Comp-Sci Grad comes onboard to his first ever 'real' job and yet .. has no idea how to even find the compiler on his system, let alone build one himself.
There is a deleterious effect with the computing industry, which happens whenever the New Cool Shit™ hits the scene - everyone forgets the Old Cool Shit™ as fast as they can - but, inevitably, eventually, you are gonna have to learn how to write a compiler, kids, and build a VM/ecosystem around it, at the very least. (Or just use Lua, lol...)
When I first encountered TempleOS, I wasn't as impressed as my peers - this is just how we did things in the 80's - but what was impressive was the persistence he exhibited in extending TempleOS in spite of the great personal suffering under which he clearly labored ..
TempleOS is just about the worse possible example of this.