Programming thread

[779777]

Terry was very smart and his dedication was seriously impressive but he's mostly liked for his personality and as others have said his insight into computer science is really valuable. I haven't actually read any of his code but I have seen a lot of him programming in his archived livestreams. Whether or not the OS is useful or shows his true genius etc is not really the point, it's about the spiritual war not the technical one

 

[Shalalallah Akbar]

Pardon the ignorance but it takes that amount of time on average to build an OS by oneself?

Linus Torvalds wrote the first version of Linux in a couple of years (IIRC for his bachelor's thesis) but that was a lot smaller in scope (being only the kernel and some drivers rather than a kernel, the GUI system, drivers and several applications including a compiler for his version of C as Terry did), and he had inspiration from Minix. Otherwise, pretty much no-one really decides to write an OS themselves (other than perhaps as a learning exercise - so basically just a kernel rather than an entire OS), so I don't think there's much to compare it against.

 

[Celestial Peace]

I'm only asking these questions just to have an idea of what it takes to write an OS (specially alone including writing your own kernel, drives and stuff).

I'm one of these folks that tried to learn programming but failed (reasons have to do with lack of enthusiasm and real hiring chances where I'm from) but perhaps this is something that'll learn better as a hobby so as to not give in to pressure.

But there's something that still keeps me from doing it and it's the feeling of "mysticism" (as mentioned by @306h4Ge5eJUJ) around tech and coding so I think it'd be better if I talk to people who either work on the field or know how to handle computers and tech.

Anyway thanks for the clarifications so far.

 

[ADHD Mate]

My project will be shutting down at the end of q1/q2.
Wonder if my software house overlords will find anything interesting, given current state of industry, or I will end up again in some old crusty maintenance project.
I am so tired of cloud bros. I would even give up home office if it meant interesting project.

 

[Rusty Metal Skull Gun]

So for those who have a knowledge on building a OS from the ground up I'll ask your opinion: Was Terry Davis that much of a genius as his fanbase (and this forum apparently) claim to be?

After talking a look at Temple OS and its features dosen't seem anything impressive and from an user POV there are some stuff that seem rather counterproductive or outdated. But the question that begs to be asked is: What was the purpose of Temple OS objectively speaking?

Well, things don't necessarily need a purpose. But I think he'd probably agree that God wants us to use our intellects, and if Terry encouraged other people in their endeavours that's a good thing.

If I had to criticize him, it would be for making excuses for avoiding the difficult bits. He only had a resolution of 640x480 because God said so. Ditto for the lack of USB. His filesystem didn't support extending files. I think his HDD driver lacked DMA support too.
But, still impressive for a one-man project.

 

[macrodegenerate]

I'm only asking these questions just to have an idea of what it takes to write an OS (specially alone including writing your own kernel, drives and stuff).

I'm one of these folks that tried to learn programming but failed (reasons have to do with lack of enthusiasm and real hiring chances where I'm from) but perhaps this is something that'll learn better as a hobby so as to not give in to pressure.

But there's something that still keeps me from doing it and it's the feeling of "mysticism" (as mentioned by @306h4Ge5eJUJ) around tech and coding so I think it'd be better if I talk to people who either work on the field or know how to handle computers and tech.

Anyway thanks for the clarifications so far.

The thing alot of people skip over is that Davis had a bachelor's degree in computer science from the 90s. Two of the greatest achievements of the OS are the Holy C compiler and the operating system itself. The basics of how to code these things and how they work are taught in CS programs today as simple, but requiring hard work.

A competent graduate today would be able to build a compiler like holy C using YACC. Operating systems are taught as well specifically memory loading/unloading.

People today think these things are easy because the tools exist now to do them. Terry didn't have these tools or easy access to knowledge on how to build an os when he made Temple OS. He just did it.

 

[779777]

Most "new" languages these days are just yacc/bison glued into llvm, shameful

 

[analrapist]

Two of the greatest achievements of the OS are the Holy C compiler and the operating system itself. The basics of how to code these things and how they work are taught in CS programs today as simple, but requiring hard work.

The integration between the OS and the compiler is another remarkable point of interest. I wish I remembered more details to regale y'all with, but it's been a long time since I delved the details. HolyC source files on TempleOS don't behave like C source files on Linux. There's a lot more streamlining of the automation. The compiler acts like a package manager in some regards.

 

[y a t s]

Terry was very smart and his dedication was seriously impressive but he's mostly liked for his personality and as others have said his insight into computer science is really valuable. I haven't actually read any of his code but I have seen a lot of him programming in his archived livestreams. Whether or not the OS is useful or shows his true genius etc is not really the point, it's about the spiritual war not the technical one

Just man up and learn what malloc and free are

This man was a gem.

 

[779777]

The integration between the OS and the compiler is another remarkable point of interest. I wish I remembered more details to regale y'all with, but it's been a long time since I delved the details. HolyC source files on TempleOS don't behave like C source files on Linux. There's a lot more streamlining of the automation. The compiler acts like a package manager in some regards.

Crunklord could tell you all the intimate details but from what I remember the compiler outputs either shared or static objects and they're loaded directly into memory, so there is no linker step instead you include the holyc file directly which will compile and load the includes as need, I assume all objects are shared globally so there is no duplicate code.

 

[analrapist]

Crunklord could tell you all the intimate details but from what I remember the compiler outputs either shared or static objects and they're loaded directly into memory, so there is no linker step instead you include the holyc file directly which will compile and load the includes as need, I assume all objects are shared globally so there is no duplicate code.

Yeah, Terry goes over that briefly in the video you linked. Apparently, compiling also extracts documentation out of the code for the inline help system. That's the kind of tight integration I was recalling. Building the system brings all kinds of additional metadata in that you just don't see on other platforms, especially proprietary ones.

 

[ADHD Mate]

Yeah, Terry goes over that briefly in the video you linked. Apparently, compiling also extracts documentation out of the code for the inline help system. That's the kind of tight integration I was recalling. Building the system brings all kinds of additional metadata in that you just don't see on other platforms, especially proprietary ones.

Would that be the present if Xerox exec weren't retarded and Smalltalk won?

 

[analrapist]

Would that be the present if Xerox exec weren't retarded and Smalltalk won?

Not sure what alternate history you're reading where Smalltalk had any kind of chance, but if LISP didn't grow past its academic niche, Smalltalk's got even less potential.

Supposing, somehow, that Smalltalk became the predominant language instead of C, today still looks like Smalltalk.NET or some such.

 

[ADHD Mate]

Not sure what alternate history you're reading where Smalltalk had any kind of chance, but if LISP didn't grow past its academic niche, Smalltalk's got even less potential.

Haven't Xerox at that time had like enough capital to become biggest player if they entered Computer market? But decided against it, as it would damage they core market?
They had the tech at Xerox PARC for sure.

 

[Celestial Peace]

People today think these things are easy because the tools exist now to do them. Terry didn't have these tools or easy access to knowledge on how to build an os when he made Temple OS.

We could argue that when he finished TempleOS operational systems were at a more advanced stage than TempleOS was. I ain't gonna argue about Windows XP and macOS because they're corporate products developed by a team but there were lots of Linux distros. What I'm saying is he could've used these tools that already existed at the time to improve his OS.

Furthermore we have SerenityOS which was also developed by only one developer but he dosen't have the same cult following as Davis has.

 

[779777]

What I'm saying is he could've used these tools that already existed at the time to improve his OS.

If you watch the video I linked he explains the difference between his operating system and moderns ones like Linux, using whatever tools isn't in line with what the operating system is trying to be, that's why the whole os and compiler are under 100k lines (the Linux kernel alone is 35 million)

 

[Belisarius Cawl]

Haven't Xerox at that time had like enough capital to become biggest player if they entered Computer market? But decided against it, as it would damage they core market?

There are a number of reasons people come up with that Smalltalk didn't have as huge of a direct impact as it might have otherwise (as opposed to indirect impact, which is actually massive). It's true that Windows and Mac OS did a shallow copy of the user interface developed at Xerox PARC and didn't implement Alan C. Kay's full vision but I really don't think it was possible on consumer-grade hardware of the time. (Remember that a number of famous games were written in assembly into the 90s.) Another reason put forward is that distribution of the full source code is part of that vision and probably required for it. That's not a problem for open source developers but it sure as hell was a concern for commercial devs of the time! I would say for me the drawback I noticed immediately after trying out Pharo (successor to another open source implementation, Squeak, that doesn't look like a toy for preschoolers) is how insular the whole system is. It's very committed to its own VM/VI model, wanting to act like a fully self-contained operating system, and things that are trivial when using Unix tools and the Unix filesystem I just couldn't figure out. I forget who made this comparison originally, but it's an apt one: the Smalltalk hot air balloon drifting away from every-fucking-thing else:

Having said that, if you do try out Pharo, be sure to make use of the wonderful method finder:

https://news.ycombinator.com/item?id=23893859

I mainly use Vim or Neovim (another reason getting into Pharo was difficult) but I'm not sure even the big name IDEs have functionality like what's shown in the above.

If you watch the video I linked he explains the difference between his operating system and moderns ones like Linux, using whatever tools isn't in line with what the operating system is trying to be, that's why the whole os and compiler are under 100k lines (the Linux kernel alone is 35 million)

People in the thread keep forgetting that the design of TempleOS was (at least according to Terry) dictated to him by Mr. God and Mr. God did not want another niggermonkey Unix clone

 

[Email]

I'm under the impression others experience the same thing sometimes,

I assume I made a complex mistake.
I see I made the complex part correctly.
I realize I made a stupid mistake.

 

[779777]

I'm under the impression others experience the same thing sometimes,

I once spent 12 hours debugging my Java code because I had accidently set a bool to default to true instead of false

 

[GhastlyGhost]

I once spent 12 hours debugging my Java code because I had accidently set a bool to default to true instead of false

The amount of times I've accidently forgotten to add results to a list after running calculations...