"Windows, Linux, Mac OS X, and the BSDs we have today are still built using the architecture developed when computing resources were limited even though it is no longer technically necessary."<p>Nonsense! Why would you waste any resources on infrastructure when you could use them for applications. Every cycle that your garbage collector runs is a cycle that the cpu could have been sleeping and saving battery life. This concept that we don't need to care about resources anymore is absolute poppycock.
The article is full of gold:<p>- OS's were designed for limited resources back in time. Now it's time to design something for unlimited resources.<p>- Attempting to retrofit a new security model to a current OS would require everything would be rewritten so we decided to rewrite an OS, which will require everything to be rewritten for it, including the OS this time.<p>- Sandboxing, local-scope FS access, user controlled permissions have never been done before, so we're doing those.<p>- Runtime analysis will give out better performance<p>- Compilers have always been hindered by OS design, so our new OS will break the barriers of compilers (finally! )<p>(and more!)
Very interesting project, Thomas. Here's one for inspiration:<p><a href="http://www.symbolics-dks.com/Genera-why-1.htm" rel="nofollow">http://www.symbolics-dks.com/Genera-why-1.htm</a><p>About time we have something modern that can do everything in that ad that's relevant today. I want my live-updating, live-fixing, per-function compiling, whole-system debugging, crash-resistant OS. If they could do it in the 80's then... you know... ;)
Does the language-based security model of 3L mean that all programs must be written in Lisp? Single-language systems tend to miss out on a lot of cross-platform software (which is typically written in or bootstrapped using C).
So OSes designed in the 70s are bad[1] but we'll use a language designed in the 50s because it's good.<p>Old stuff is bad except when we like it, I guess.<p>[1] And quite dishonest to lump WinNT in with that...
I have always been curious about Lisps but the lack of static typing turns me off. Sure you can have a type system (typed racket) but it doesn't feel organic, also then why not use any other statically typed language (haskell, scala, ..). With all the hype (very likely well founded) surrounding Lisps, I feel like I am missing out.
Too much network and filesystem data moves through an OS before the user application can work with it. As data bandwidth increases, this becomes a bottleneck. It would be great if an OS could push this out to userspace and focus on zero-copy stacks for network and disk.
I know (with software) to never say never, but to be quite frank, I'm content with the current Unix paradigm, and I haven't seen a compelling reason why any new OS would be fundamentally superior.<p>But I would absolutely love to be proven wrong...<p><a href="http://www.catb.org/esr/writings/taoup/html/plan9.html" rel="nofollow">http://www.catb.org/esr/writings/taoup/html/plan9.html</a>
"Another change since the original OSes were initially being built is the need for security"<p>I do not know what's referred to as "original OSes" but many old operating systems had actually pretty decent security architecture. MULTICS and VMS, for example, and many others too.
Keep going!
To make a new OS successful, there are too many difficulties. Existing competitors, acceptance of consumers and cost of shifting from the old to the new. We are happy to see a new OS. Hope it can go well!
There is a market for a new OS, but security alone is not enough reason to make people switch.
Supporting Unicode would affect users a lot more.
I wish I saw an OS with good support for Chinese.
Fascinating project.<p>I'm glad that alongside the philosophy of this being a FOSS endeavor, there's the practical aspect of this being adjustable on the fly by users in the old Lisp machine tradition.<p>Though I'm sure it's too early for outsiders to shape the PoC code, is it available anywhere for us all to follow along?
Entirely new OS would be a waste of resources, but some big player could donate few millions to boost the development of Wayland. What the markets lacks is a system with openness / hardware support of Linux combined with solid graphical layer of OS X. Or maybe AOSP could transform into true open source (culturally) project?
When I read this I raised my eyebrow. I hardly think that Lisp (1958) can be the base of the operating system of the future. Their texts are full of bold statements about how easy and natural lisp is and how everything else sucks. Nevertheless it's interesting to look at their references:<p><a href="http://3lproject.org/references" rel="nofollow">http://3lproject.org/references</a><p>I understand than in short, the security model is about how to make sure a process does not do more than what it's supposed to. In this case it's lisp environments that provide the necessary sandboxing.