5 ms·

A serious question: When will we see a truly new OS built from the ground up? If there is a space where there seems to be little innovation it is in the space o

A serious question: When will we see a truly new OS built from the ground up? If there is a space where there seems to be little innovation it is in the space of new OS paradigms. Flavors of Windows or Linux - are there really any commercially ready alternatives today or in the very near future?

michaelmior 11y ago
New operating systems are still being created. Barrelfish[0] for example is quite promising. The "commercially ready" part is much more challenging. [0] http://www.barrelfish.org/ http://www.barrelfish.org/
- iso8859-1 11y ago
  Genode has commercial support: http://genode.org/ http://genode.org/
ferrari8608 11y ago
Haiku comes to mind. It's an open source reimplementation of BeOS. Unfortunately it's still alpha like ReactOS.
ant6n 11y ago
How about an operating system that doesn't execute native code directly anymore. No Segfaults. All programs live in the same VM space.
- cpitman 11y ago
  It's pretty much dead now, but Microsoft Research had a project called Singularity (http://research.microsoft.com/en-us/projects/singularity/ http://research.microsoft.com/en-us/projects/singularity/). It's an entire OS implemented in Spec#, which is a subset of C# + Code Contracts. It was built so that contracts between programs could be statically verified when compiled, allowing applications to all run in the kernel's address space.
- ferrari8608 11y ago
  Isn't that kind of what Android does?
  - ant6n 11y ago
    Well, if the JVM was the basic layer of the operating system, then yes. But really it's some form of Linux, there's a lot of native code running, system calls, user and kernel mode etc. Apps can use native code too.
0xcde4c3db 11y ago
Serious OS research has tended to get stuck in academia and research labs for the past ~20 years; probably something to do with the worse-is-better pricnciple. L4 [1] has managed to sneak into a lot of phones in basebands and "security processors", though. [1] http://www.l4hq.org/ http://www.l4hq.org/
iopq 11y ago
There's an OS written in Rust: http://www.redox-os.org/news/this-week-in-redox-8/ http://www.redox-os.org/news/this-week-in-redox-8/
pjc50 11y ago
The last successful launch of a "new" operating system was maybe iOS? Whereas there are plenty of expensively developed commercial failures littering the path. A serious problem: who would pay for the development of a new OS, or sink a very large chunk of free time into it? What properties would it have that could not be achieved starting from one of the POSIX OSs? Who would port all the apps? Why would people - end users - adopt it?

michaelmior 11y ago

New operating systems are still being created. Barrelfish[0] for example is quite promising. The "commercially ready" part is much more challenging. [0] http://www.barrelfish.org/ http://www.barrelfish.org/

iso8859-1 11y ago
Genode has commercial support: http://genode.org/ http://genode.org/

ferrari8608 11y ago

Haiku comes to mind. It's an open source reimplementation of BeOS. Unfortunately it's still alpha like ReactOS.

ant6n 11y ago

How about an operating system that doesn't execute native code directly anymore. No Segfaults. All programs live in the same VM space.

cpitman 11y ago
It's pretty much dead now, but Microsoft Research had a project called Singularity (http://research.microsoft.com/en-us/projects/singularity/ http://research.microsoft.com/en-us/projects/singularity/). It's an entire OS implemented in Spec#, which is a subset of C# + Code Contracts. It was built so that contracts between programs could be statically verified when compiled, allowing applications to all run in the kernel's address space.
ferrari8608 11y ago
Isn't that kind of what Android does?
- ant6n 11y ago
  Well, if the JVM was the basic layer of the operating system, then yes. But really it's some form of Linux, there's a lot of native code running, system calls, user and kernel mode etc. Apps can use native code too.

0xcde4c3db 11y ago

Serious OS research has tended to get stuck in academia and research labs for the past ~20 years; probably something to do with the worse-is-better pricnciple. L4 [1] has managed to sneak into a lot of phones in basebands and "security processors", though. [1] http://www.l4hq.org/ http://www.l4hq.org/

iopq 11y ago

There's an OS written in Rust: http://www.redox-os.org/news/this-week-in-redox-8/ http://www.redox-os.org/news/this-week-in-redox-8/

pjc50 11y ago

The last successful launch of a "new" operating system was maybe iOS? Whereas there are plenty of expensively developed commercial failures littering the path. A serious problem: who would pay for the development of a new OS, or sink a very large chunk of free time into it? What properties would it have that could not be achieved starting from one of the POSIX OSs? Who would port all the apps? Why would people - end users - adopt it?