A Technical Look Inside TempleOS

jones_supa writes: TempleOS has become somewhat of a legend in the operating system community. Its sole author, Terry A. Davis, is a special kind of person, who has a tendency to appear in various places with a burst of strange comments. Nevertheless, he has spent the past 12 years creating a new operating system from scratch, and has shipped a functional product. An article takes a constructive technical look at the internals of TempleOS: installation, shell, file explorer, hypertext system, custom HolyC programming language, and interaction with hardware. The OS ships with a suite of several tools and demos as well. To see the sheer amount of content that's been written here over the years, to see such effort expended on a labor of love, is wonderfully heart-warming. In many ways TempleOS seems similar to systems such as the Xerox Alto, Oberon, and Plan 9; an all-inclusive system that blurs the lines between programs and documents.

Comments

Interesting person

[spiritplumber]

This guy hangs out on hackernews, he's... well frankly he's a bit of a religious nut, but he doesn't preach at you or anything (unless you ask). Definitely a work of passion.

Re:Interesting person

[LWATCDR]

Not really. I know lots of liberals that are very racist but in a very dishonest way. They will go march in a MLK day parade but send their children to all white private schools and only have friends that are upper middle class and usually white.
While I know a good number of people that are conservative that have all sorts of friends.
So you are labeling broad categories of people with attributes that are not intrinsic to their stated belief systems.

Re:Interesting person

[Bacon Bits]

Obviously.

Like TFA says, you need to look at it like a research OS. One that has critical flaws that are designed in, but that design has a purpose: to function in a different way. Doing so can expose new lines of thinking and novel approaches to "solved" problems. No, it won't function in a real world of networked computing, but it's not supposed to fit into that idiom. It still has some very interesting ideas. Spending time solving every problem again isn't the goal. There was a time when virtual memory didn't exist, for example, and computing still worked well enough to run businesses, banks, telephone networks, and governments. We don't need a research OS to show us that virtual address spaces are useful. We know that already. So, ignore it as not relevant, and do something that is a novel approach.

Looking at the ideas in TempleOS like they will replace Linux or Windows is silly, but they might give us ideas for new types of computing. The idea that everything would be better as a Linux device is, quite honestly, poisonous to the development real progress in the field of computing as a whole.

Maybe TempleOS is like non-Euclidean geometry. Sometimes need drives the development of new math -- Newton's development of Calculus -- and sometimes the math is developed and sits idle, doing nothing for nearly a hundred years before changing the world -- like Boolean algebra. A computer system is just a very complicated set of mathematical rules. Changing the rules of math and seeing what happens has been one of the major forces of change, as different systems are often best expressed in different forms of mathematics.

Does TempleOS make it easier to understand how computers operate? Does it make it easier to learn what a program actually is? Is it just an example of being closer to the bare metal, like you were flipping bit switches on an old Altair 8800? What if, for example, a system like this makes it very easy to model artificial intelligence? It can basically reprogram itself, after all, as everything is JIT and source code is readily available for literally everything at all times. That seems incredibly powerful. Is it possible to write a self-refining program in HolyC?

Just because I don't understand what something could be used to do doesn't mean it's useless. It might just mean I don't have a very good understanding yet. The questions to ask are, "What kind of system benefits from the design of TempleOS? What kind of system benefits from raw, unimpeded access from the user or input to the hardware?"

Hmm...

[EmeraldBot]

TempleOS has always struck me very similar to the ancient LISP machines, the ones that everyone loves so much. It's such a shame that the OS will forever be held back by its author, as well as some of its more practical limitations (*cough* no sound support *cough*), because it has some very good ideas. It particular, the indexing and documentation system are just overall fantastic; Java is widely lauded for its excellent documentation features, but it doesn't have anything compared to this. The shell is another really awesome idea; a multimedia shell is something that I've actually never considered, to be totally honest, it never crossed my mind. Imagine a shell you could just live in; one in which you could browse your system, listen to music, do your email, etc. all without ever having to leave your coding environment. I know emacs exists, but it's not quite on this level - I wish other operating systems like FreeBSD or Linux had an equivalent.

Unified Hypertext is impressive (among others)

[CrashNBrn]

I saw TempleOS on BetaNews -- at least 5 years ago now. Never had the time nor inclination to see what it was all about, as it looked just a little bit too Old-School (the EGA/VGA 16 color). This "technical look" is actually pretty interesting, especially the way the system is built on and uses hyperlinks, the file format is described as

HTML, JSON, XML, shell scripts, source files, text files – TempleOS replaces all of these via one unified hypertext representation.

Maybe the WhatWG or W3C could learn a thing or three.