Porting Inferno OS to Raspberry Pi | Hacker News Reader

Porting Inferno OS to Raspberry Pi

github.com

68 pointsrcarmo15 comments

You can also run inferno on a raspberry pi pico. I did the port here https://github.com/caerwynj/inferno-os/tree/pico

Interesting. I thought Inferno needed > 1MB ram.

:-O do you have some demo of it running or a blog post about it. I'd love to learn more.

I wish Inferno/Plan 9 had caught on more, but sadly I think it's not quite the abstraction(s) we need in the modern day.

Modern-day programmers reinvent Plan 9 abstractions every few years, poorly.

It did caught, as a research system: it is/was a fertile ground, from which many ideas made their ways to mainstream systems. There's so much inertia outside of the Plan9 realm that it essentially prevents any potential challengers: the threshold is high.

I do think Plan9 did succeed to the best of its abilities, given the context.

This comment makes me sad. Plan 9 is still very much alive - see 9front.org a fork that is actively maintained.

> but sadly I think it's not quite the abstraction(s) we need in the modern day.

This leads me to believe you formed this opinion without ever exploring it. Plan 9 is built for a networked world like the one we live in. Perhaps you should study it a bit more before drawing such conclusions.

I'd suggest you check out the adventuresin9 youtube channel.

> I'd suggest you check out the adventuresin9 youtube channel.

Unless I'm mistaken, there is no way to watch YouTube on Plan 9, which rather reinforces the point that it is not made for the modern world, no?

This has less to do with the technical merits of Plan 9 and more to do with the fact that nobody has ported a modern web browser to Plan 9 (there are some browsers, but none with the feature parity of Firefox or Chrome). A slightly easier route to YouTube support for Plan 9 would be if somebody ported youtube-dl, but this work is also non-trivial.

It’s the classic chicken-and-egg problem of adopting new platforms: the platform needs software to attract users, but developers won’t develop unless there are users or there’s a high likelihood users will come. It’s not that Plan 9 isn’t made for our world; it’s just that Linux/Windows/macOS/BSD and their software ecosystems are good enough for many users, and those who want to take advantage of Plan 9’s features have to contend with a lack of software in many domains. It’s similar to the reason I code in Python for my machine learning job despite the fact I’d prefer a Lisp or a statically-typed functional programming language: Python has a far richer ecosystem of machine learning and data science libraries.

http://wiki.9front.org/youtube

If that is unsatisfactory then you can either A. use vmx(1) to run Linux and run a browser in there. or B. write code and contribute instead of complaining.

Though what does youtube have to do with modern computing? It's just an application. I'm talking about the underlying architecture.

Actually, there is (I spotted a YouTube downloader someplace), but besides that "point", the videos in that channel are _very_ interesting if you care anything about modern microprocessors (he's ported Plan9 to quite a few devices).

I did. I really wanted to like Plan 9, but I fail to see how it will ever be more than anything than a research system for playing around with. I read the Plan 9 Architecture paper, deployed to VMs (because that's about the only thing it will run on), and gave it's design decisions serious consideration. The virtual union filesystem I think has merit, but it's not without tradeoffs. If we want to see Plan 9 in any form used in serious, I think it would have to be as a virtual userspace implemented on top of Linux with more serious thought given to performance, error handling, and introspection.

> (because that's about the only thing it will run on)

Sounds like you aren't running 9front. 9front runs on lots of things. We just had a hackathon in early august and there were numerous PC's running 9front. adventuresin9 is working on a pine phone port along with a bunch of mips and arm devices. You aren't paying attention to what is important and instead getting mired in the tediousness of applications.

> I think it would have to be as a virtual userspace implemented on top of Linux with more serious thought given to performance, error handling, and introspection.

There is no point in that at all. The advantage of plan 9 is plan 9's architecture - its not an application. Might as well use p9p on Linux with a rio theme. The only issue in terms of performance is 9p latency over distances but these are fixable problems. But over a LAN its not an issue. Patches welcome. Lastly: "error handling, and introspection" - you need to clarify these as this just reads as fluff.

This port is ancient history.

Have a look at: https://dboddie.gitlab.io/inferno-diary/index.html

That's the kind of blog I wouldn't mind reading if it had an RSS feed.

Reply on news.ycombinator.com