Comment by randyfox

Microsoft came close with Midori but bailed out and canned the product just before it should have been released in alfa / beta 1

> Midori is an experimental managed code operating system that was in development until 2015. A joint effort by Microsoft and Microsoft Research, it had been reported to be a possible commercial implementation of the OS Singularity, a research project begun in 2003 to build a highly dependable OS whose kernel, device drivers, and application software would all be written in managed code. It was designed for concurrency, and would run a program spread across multiple nodes at once.[1]

[1] https://en.wikipedia.org/wiki/Midori_(operating_system)

https://joeduffyblog.com/2015/11/03/blogging-about-midori/

Because writing even a remotely modern OS is really really hard. I speak from experience. Even getting old hardware from the 90s and early 00s to work is a pain. Then of course you have the more modern standards (although even SATA is still modern for some definition of modern, but AHCI is a nightmare), and then you have things like modern NICs and GPUs which has documentation that is very hard to find, or in the case of even modern Intel GPUs, documentation that is 13 plus volumes and is absolutely massive... And the list just goes on and on and on. Before you know it your codebase is 100k LoC and like 80-90 percent of it is device drivers alone. And if you thought all that was bad, wait until you get into ACPI...

Yeah. OS dev is, I think it's safe to say, the hardest and most difficult project a software engineer could do, right alongside a modern compiler if you ditched LLVM and decided to make your own backend.

Because many open an UNIX book, and rather copy what was already done.

Note that Smalltalk, Interlisp-D, Mesa, Cedar, safe systems programming are also 60/70's tech, but these ones hardly anyone bothers to copy in such attempts.

My dream is to work on an operating system that at least gets us to the 1990s and 2000s when it comes to research ideas.

I have a soft spot for the Smalltalk-80 environment and Lisp machines. They had a single address space. In my opinion, the two most interesting things about these environments are (1) their facilitation for component-based software based on live, dynamic objects, and (2) the malleability of the system, where every aspect of the system can be modified by the user in real time.

Of course, a critical downside of Smalltalk-80 and Lisp machine environments is the lack of security; any piece of code can modify the system. There are two solutions to this that I'm thinking about: (1) capability-based security for objects in the system, and (2) work on single-address space operating systems that still have memory protection (Opal was a research system that had this design; see Sharing and Protection in a Single-Address-Space Operating System [Chase et al. 1994]).

One of the nice things about Lisp is its metaprogramming facilities, from macros to the metaobject protocol. Metaprogramming makes it feasible to implement domain-specific languages that make expressing problems more aligned to their domains.

During the late 2000s and early 2010s, Alan Kay's Viewpoints Research Institute had a project named STEPS that investigated the pervasive use of DSLs to implement an entire desktop environment. They did not use Lisp as a substrate, but they did use OMeta (https://tinlizzie.org/ometa/) for handling parsing expression grammars (PEGs), which are used to describe many of the systems in STEPS. Two DSLs that immediately come to mind are one for describing the 2D graphics system and another for describing TCP.

So now I've described my dream substrate: a single-address operating system with capability-based security, where each subsystem is expressed as a live object, ideally coded in a DSL.

Now comes the interface. The programmer's interface would be similar to Smalltalk-80 and Lisp machines, with a live REPL for interactive coding. All objects can be accessed programmatically by sending messages to them. The end-user interface would be heavily based on the classic Mac OS, and applications would conform to human interface guidelines similar to System 7.5, but with some updates to reflect usage patterns and lessons in UI/UX that weren't known at the time. Application software would be similar to the OpenDoc vision, where components can be combined based on the user's wishes.

The end result sounds like a synthesis of various Apple projects from the late 1980s until 1996: component-based applications backed by a live object system with capability-based security.

This is my dream and is a side project I'd love to create.

I would argue that iPadOS (built on iPhone's coat tails) moved the field forward significantly in terms of isolation and user security.

While this has left a long tail of inconveniences, many resolved and some not, I am very confident that using 1 app on my iPhone/iPad will not leak data to another in any case that I am likely to care about as a non-significantly interesting person (political figure, etc).

... and for those people Apple even makes lockdown mode to move the bar, while acknowledging it adds extra inconvenience: https://support.apple.com/en-au/105120

I have no such confidence about macOS, Linux or Windows, in fact the reverse. macOS has done the best at trying to bolt on some sandboxing (and linux has it too) but that's still very holey and not all-in like iOS/iPadOS has ended up.

Yes, I know there have been many bugs and leaks in iOS but the security level is far and above the desktops currently, and designed that way from the ground up. So when they finally make something work like copy and paste or sharing between apps, etc... it's by and large done very well.

It's been very difficult to add that kind of thing to Linux because you're trying to do the reverse and lock things down and it breaks everything... making it very challenging.. as opposed to Apple where basically nothing useful worked at the start (no copy/paste, one app at a time, no meaningful filesystem, etc).. but managed to get the product successful in the limited state and has slowly unlocked that stuff over time. Admittedly very slowly.

I cannot speak for Android as I just have never used it or surrounded myself in info about it's design, security, etc.. it may well be very similar although they from my casual observation seemed to do a much worse job at granular privacy permissions (e.g. for the longest time permissions were all granted at install time, and so many apps want so many most people are blind to it.. as opposed to Apple's model where even if notarised for something on the app store in most cases you have to agree to it when the app first uses it.. I know they fixed that a while back but I have no idea how well things have transitioned to that now). As a very techy person deeply knowledge in many things, and using desktop Linux since 2002, it's kindof a hilarious personal failing that I have never used Android.. I really should try and resolve that at some point.

How about the immutable OSes like Fedora Silverblue

It takes an astronomical amount of work to not only write an OS but handle 30 years worth of both hardware, and lessons learned.

People do write new OSes from scratch all the time, you can see the result of no popularity/usage because they can't do much for even enthusiasts.

Not to mention, if you want users, you need apps, and that involves trying to run what already exists or making it as easy as possible to port. All of that shapes the OS.