Good opportunity for a free pascal operating system...

[MarkMLl]

If I had that level of proficiency I'd budget a couple of weeks to design and build a custom CPU to go with it.

Oh, that, that! I support it with both hands and feet In this project I will support you... spiritually Because unfortunately I can't design CPU and I don't have the right budget

I was, of course, trolling OP there. However if I were getting back into this sort of thing I would- seriously- consider a custom CPU, probably RISC-V with a segmenting MMU.

Segments got a bad name because of some fundamental mistakes that Intel made with the '286 and its derivatives. Unfortunately, people focus on the wrong thing and assume that The Problem was the 64K maximum segment size: if that was all there was then segments should have survived in usable form once the '386 allowed them to be 4Gb.

The real problem wasn't the size of the segments, but the total number: 8K in the LDT and 8K in the GDT if memory serves me correctly. Once you've got a development system with each unprivileged read-execute segment being shadowed by a privileged read-write one for debugging, and once you've got every data structure that might be moved around in its own segment with its own descriptor the whole thing becomes untenable very quickly.

I think there's ways around that problem which would bring back the level of hardware debugging support that made system design "a pleasure" on the '286 or '386, while at the same time ensuring adequate performance by dedicating some of the vast amount of cache space on a modern chip to address caching. However since this is moving somewhat off topic I'll stop there...

...with a final note that Zynq doesn't allow the FPGA to be interposed between the CPUs and memory, and I'm unsure about CHERI.

MarkMLl

[Joanna from IRC]

Has everyone been watching what is happening to the linux kernel. It seems linux will go down the drain. This is a good opportunity fore the free pascal community to do a operating system that learns from the mistakes of the linux foundation. Would love to hear what the community thinks about these events.

https://www.youtube.com/watch?v=xPF2NSFvNCc

Good idea , please keep us posted on your progress of creating a new operating system using fpc. 

[cpicanco]

Would love to hear what the community thinks about these events.

My political thought was not well received. I respect that, it is completely understandable.

What I am doing now is thinking "technically" about the events that motivated this discussion. As you mentioned, writing a general-purpose operating system comparable to GNU/Linux requires many human resources, decades of work, and stability. Since it is a long-term task, it is essential to consider beforehand the some factors that can interfere with the development of the system, especially supposedly general systems. Questions like "Who will use the system?" and "For what purpose?" are also essential for planning the logistics of such an undertaking. Will the system be used by people in the public sector or the private sector? Both? Will it run data management programs in schools, hospitals, etc.? Would the purpose be financial management in the financial sector? Would the system have applications in war, in a military context? To whom would it offer power and advantages? Will you be proud of the system you helped write when it is ready and fulfills its purpose? Was the system made to educate or to kill children in schools? Will it serve soldiers on the battlefield or hospitals? Both? Is the writing of this system purely for educational purposes? After the system is ready, is it possible to estimate its longevity? What should be part of the planning to ensure that there will be educated people to maintain it when it is ready? Under what circumstances is it possible to plan an operating system without thinking about politics?

[nomad]

Would love to hear what the community thinks about these events.

My political thought was not well received. I respect that, it is completely understandable.

What I am doing now is thinking "technically" about the events that motivated this discussion. As you mentioned, writing a general-purpose operating system comparable to GNU/Linux requires many human resources, decades of work, and stability. Since it is a long-term task, it is essential to consider beforehand the some factors that can interfere with the development of the system, especially supposedly general systems. Questions like "Who will use the system?" and "For what purpose?" are also essential for planning the logistics of such an undertaking. Will the system be used by people in the public sector or the private sector? Both? Will it run data management programs in schools, hospitals, etc.? Would the purpose be financial management in the financial sector? Would the system have applications in war, in a military context? To whom would it offer power and advantages? Will you be proud of the system you helped write when it is ready and fulfills its purpose? Was the system made to educate or to kill children in schools? Will it serve soldiers on the battlefield or hospitals? Both? Is the writing of this system purely for educational purposes? After the system is ready, is it possible to estimate its longevity? What should be part of the planning to ensure that there will be educated people to maintain it when it is ready? Under what circumstances is it possible to plan an operating system without thinking about politics?

General purpose means it can be used by everyone. Free pascal has been around for a long time and developed. I think it would be nice to have lazarus boot up in a operating system that is compiled in free pascal. Seeing the comments makes me wonder if many of you actually believe free pascal is a good language like c and c++. I guess you do not have the confidence in the free pascal language. Even linux kernel developers is willing to considering rewriting its kernel in rust. I think I said enough.

[LV]

Linux began in 1991 as a personal project by Finnish student Linus Torvalds to create a new free operating system kernel...

...fortune favors the bold!

[Awkward]

Linux began in 1991 as a personal project by Finnish student Linus Torvalds to create a new free operating system kernel...

...fortune favors the bold!

But it based on UNIX experience and use C language which was used for lot of program code (by many tools) already

[LV]

Almost forgotten the Apple Pascal Operating System?

[MarkMLl]

I think I said enough.

Quite frankly we agree.

It's not that it wouldn't be a fun project: perhaps 50 man-years of work taking a minimum of five years... during which time multiple participants would die or leave for other reasons.

For that reason, it's totally unrealistic. If nothing else, there are far more important things to do which are directly relevant to the survival of Pascal as a language.

MarkMLl

[dbannon]

Come on Nomad, I gave you a hint further up, if you want to poke your finger in Linus Torvalds' eye, all you need replace is the Kernel. Its all under git, you fork it, then replace a few units, one by one. The license permits that.

Linus did not start the Linux ball rolling by putting out a call for contributors, he started coding. And when people saw what was happening, they joined in.  Its how Open Source works.

Get a few units working and you will find it getting easier. We will all join in when we see that progress. Linus has no hand in all the GNU stuff, indeed they do not get along at all well. Focus on just the kernel !

Davo

[Awkward]

But isn't Oberon more  useful than pascal for kernel?

[MarkMLl]

But isn't Oberon more  useful than pascal for kernel?

"Pascal as defined by Wirth"... undoubtedly. But Wirth didn't really say what had to go into Pascal's system unit, didn't define a way of linking stuff written in assembler and didn't say anything at all about an RTL, so it would be easy enough to put the context-switching magic in the RTL written in assembler and carefully tailored for the target.

I don't like boring people by being too detailed about what I've done in the past, but when I was writing a microkernel for '286 PM I found that the Modula-2 coroutines mapped nicely onto the TSSes etc. without custom assembler. But even there there was RTL/startup code written IIRC in assembler that I had to tweak... I forget the detail but broadly speaking it is difficult if not impossible to avoid that sort of thing.

And even on the Burroughs mainframes "programmed only in ALGOL", there was handcrafted machinecode in the coldstart cards and at the start of any bootable disk. And their "ALGOL tweaked to write the operating system" provided a ruddy great word array representing physical memory, together with predefined intrinsics for every machine-level operation. In short the system-level code (MCP etc.) wasn't very different from the sort of thing that most people write in assembler plus macros for BEGIN, END, and structured control flow.

So I can say from experience that Modula-2 was "suitable for system programming" by virtue of having the sort of low-level capabilities that were also in Turbo Pascal by the early 1980s, and the implementations that I saw all had small assembler-written RTLs. And while I've not looked at Oberon (the language) I presume that Wirth continued along the lines that he explored with Lilith etc.: I believe that he started looking at FPGAs but that sort of thing /still/ requires manipulation outside the high-level language: even if it's an HDL such as Verilog rather than assembler code.

MarkMLl

[Handoko]

Seeing the comments makes me wonder if many of you actually believe free pascal is a good language like c and c++. I guess you do not have the confidence in the free pascal language.

It seems you misunderstood. We do not mean to discourage you, we just want to warn you what you want to do is not a simple task.

Not once or twice. There were several times, someone seemed to pop up from nowhere, told us Lazarus/Pascal was great and they had some awesome ideas to make Lazarus even better. We should do a, b, c and then adopt modern features from x, y, z, also we had to rename Lazarus to something that sounds cool. Some even personally contacted me discussing his ideas and asking me advice. He showed me a list of things he was going to do, interesting. They were full of confidence but not long later, they just vanished without trace.

They, also you forget many important things. Who is going to fund the project? I guess you know writing a new OS can't be finished in just one or 2 months. It will take years or more. We have our job, we need to make money and we have to eat. We sleep 7 hours a day, spend 7 hours working to earn money. I personally have to spend some hours for socializing with my friends, playing computer games and do reading. If you expect us to spend our precious time volunteering in your project, then sorry. But if you tell us Microsoft or Meta or Elon Musk is going to fund this project and we will be paid, I'm in. Money is not everything but money is an important thing.

Think again, as I said they were full of confidence but not long later, they just vanished without trace. Why that happened? Because they only had half baked idea. They did not have the skills, they didn't know how to manage software development, they didn't know how and where to start. I know your suggestion to write a new OS is good, that sure is a good thing that Lazarus/Pascal community can be proud of. But. Who is the project manager? How do we start? On what machine should we start from? Why? Do you already have a prototype? How to manage this multi-programmer project?

Do you know the procedures to develop software? I mean big software not just simple programs with hundred lines of code. If you can't tell me the steps of developing a new piece software, I recommend you "Code Complete". That book will explain you planning is most important part in software development. Without proper planning, it will sure to fail. I own the book, the second edition.
https://en.wikipedia.org/wiki/Code_Complete

If you're really serious want to write a new OS. We are open to hear what you've already had. Show us your prototype. What you've already done or going to do? At least, 'they' could show me a list of what he was going to do.

[Bogen85]

Got to this thread late.

To think Linux is going down the drain is laughable (no offense to anyone here intended).

Billions of USD have been dumped into Linux development over the past several decades, billions more will be in upcoming decades.

And yeah, I'd already full aware of the events surrounding the video that OP here shared the link to.

Events like are show in the video (developers losing access to Linux kernel development due to sanctions) are small hiccups to the Linux ongoing Linux development train.
A train with that kind of momentum is not slowed down by something like that.

I don't want to downplay the significance of those events (developers losing access to Linux kernel development due to sanctions) but this forum is not the really the place for that type of discussion.

Companies all over the world have products based on Linux, spanning back spanning back decades, and more so over the past 15 years.

Products (hard and soft) where the cost to customers ranges from tens of USD to hundreds of thousands USD, and even higher. And that is not slowing down. At all.
(or even less, one can rent Linux VMs for under 5 USD a month)

But for the sake of hypothetical scenarios, if Linux were to somehow falter in upcoming years, some new OS requiring man centuries (millennia?) of new development would likely not be what replaces it.

For many that need that type of high grade server OS, it would be likely be something like FreeBSD that already has a lot of momentum, predating the Linux momentum.

https://freebsdfoundation.org/netflix-case-study/

All the other points that I'd care to build on for this have already been hammered home (addressed in this topic thread) by other multiple long time forum members, so I don't need to add to those.

Many (including myself) back in the 1990s jumped on the Linux bandwagon well before the dot COM bust. Not because we were Linux fan boyz and girlz, or GNU (shudder the thought*) fan boyz and girlz, but because we saw Linux as the future of IT.

*I'm not trying to disparage the GNU project, it has played (and is still playing) a vital role in all this, but, as already noted in this thread, they are not always the easiest open source movement to work with and I'll leave it at that.

And post 2010 (even a bit before) that has more than proven to the be the case.

IBM, Google, Microsoft, Amazon, Oracle, RedHat (now IBM), Meta, and many others, Linux is at the core of what makes them the most revenue,

(spin up "Dad, What Are Clouds Made Of?" "Linux Servers, Mostly." meme...)

And is used at the core of banking and automotive industries, NASA, CERN, and industries that work with them, and many others, who at large are not even seriously considering to switching to another OS any time soon.

For me personally if Linux would somehow (I'm not seeing it in the foreseeable future) turn into a massive train wreck, I'd jump on the FreeBSD train (which I've not never fully gotten off of anyways) and would just keep plowing forward.

Yeah, there are some new OSes being written in Rust, but they would have to be proven in many industries first before I (and I'm sure many others) would even consider one of them.

[Bogen85]

Seeing the comments makes me wonder if many of you actually believe free pascal is a good language like c and c++. I guess you do not have the confidence in the free pascal language.

...

They, also you forget many important things. Who is going to fund the project? I guess you know writing a new OS can't be finished in just one or 2 months. It will take years or more. We have our job, we need to make money and we have to eat. We sleep 7 hours a day, spend 7 hours working to earn money. I personally have to spend some hours for socializing with my friends, playing computer games and do reading. If you expect us to spend our precious time volunteering in your project, then sorry. But if you tell us Microsoft or Meta or Elon Musk is going to fund this project and we will be paid, I'm in. Money is not everything but money is an important thing.

Yes! This! I'd jump on board too if that were the case!

In the mean team, using Linux and developing and supporting software and services that runs on Linux has been what has been paying my bills for the past 18+ years.

And I don't see that changing any time soon, even with the occasional news of developers leaving (or stepping down) for this an that (which has been going on in some fashion for 15+ years).

[MarkMLl]

And I don't see that changing any time soon, even with the occasional news of developers leaving (or stepping down) for this an that (which has been going on in some fashion for 15+ years).

I think it's worth noting that "the Linux kernel developers are fragmenting over Rust" has been very much overegged. The current situation is that it's being used for a very small fraction of the overall collection of device driver modules, and where there /is/ friction it appears to be more because the "old hands" resent being asked to explain or even- HORROR- /document/ their APIs.

From his slides, Almeida showed an example of how the Rust type system can eliminate certain kinds of errors. He noted that the iget_locked() function in current kernels has a complicated set of requirements. Callers must check to see if the return value is null and, if it is not, then the contents of the returned struct inode need to be checked to see if it is a new or existing inode. If it is new, it needs to be initialized before it can be used; if that fails, iget_failed() needs to be called, he said.

There was some discussion of the finer points of what callers of iget_locked() need to do, with Al Viro disagreeing with some of what Almeida had on his slide. That went back and forth, with Overstreet observing that it was exactly that kind of discussion/argument that could be avoided by encapsulating the rules into the Rust types and abstractions; the compiler will know the right thing to do.

...

Part of the problem, Ted Ts'o said, is that there is an effort to get "everyone to switch over to the religion" of Rust; that will not happen, he said, because there are 50+ different filesystems in Linux that will not be instantaneously converted. The C code will continue to be improved and if that breaks the Rust bindings, it will break the filesystems that depend on them. For the foreseeable future, the Rust bindings are a second-class citizen, he said; broken Rust bindings are a problem for the Rust-for-Linux developers and not the filesystem community at large. -- https://lwn.net/Articles/978738/

In retrospect, it is universally agreed that Ts'o overreacted. However I think it's fair to add that he should not be interpreted as accusing the Rust for Linux sub-project as "getting everyone to switch over to Rust": what he said was "switch over to the religion", not to the language.

So as far as the Linux kernel is concerned, that particular episode was a bit of a storm in a teacup. However I think it also highlights something important which impinges on the Pascal community: there are software developers who still- after 50 years of gradual improvement in most languages- are resolutely opposed to the idea that strong type handling and well-defined APIs are actually there to help people.

MarkMLl