views:

346

answers:

4

I was browsing for an open source kernel when I ran across SANOS which feels like something worth keeping alive as open source. To bring it up to modern standards, the following enhancements need to occur:

  • Support for Multiple CPUs/Cores
  • Add IPv6 (currently only supports IPv4)
  • Ability to run 8192 concurrent threads
  • Ability to support additional emulation required for JDK 1.6

So, I have the following questions:

  1. Does anyone here have experiences they would like to share using sanos?
  2. What is the most expedient way to attract others who would like to contribute time to make it better?
  3. In what ways is it better than either the kernel used in BSD and/or Linux?
  4. Is there merit in porting Ruby to this platform?
+1  A: 
  • The most expedient way to attract others would be to make it a more useful and practical OS, for example by the four enhancements mentioned. Once others see the merits of the OS they will want to develop and enhance it themselves, like you do.
  • Sanos is mostly POSIX compliant, so it is able to run most programs for BSD, Linux and other POSIX compliant OSs. However, Sanos attempts to support a minimal subset of those systems, geared towards its specific application. Hence it lends itself to both portability and speed. Its security remains relatively untested compared to BSD & Linux, however.
  • There could be merit in porting Ruby MRI to Sanos, it should be rather easy.

The most efficient way to enable JDK 1.6 would be to scrap the Win32 layer (or just Hotspot) and port Kaffe in its place. This should be relatively easy, and is easy to update. However, the best place to find out the answers to your questions would be to contact the Sanos development team and discuss these issues with them. They will tell you how you can help.

EDIT: The maintainer of Sanos has a post regarding a performance comparison with other operating systems.

Kelden Cowan
+2  A: 

I'll keep it short and simple. I don't want to rain on your parade but, commercially, I think maintaining SANOS is a waste of time.

The landscape is littered with the corpses of FOSS projects that were a good idea but turned out to be unnecessary or never gathered enough support. You may get a small cadre of people working on a project like this but to what end. Where are the installed bases for BeOS, Haiku, ReactOS et al?

Let's be clear. Linux is so far ahead in terms of development effort that those other projects would take decades of effort just to catch up. I'm not talking about technical merit here (although that may be the case as well) but penetration into the mindsets of business. And business is, unfortunately, where the money comes from.

Those four bullet items you list are already available in Linux so what would be gained by adding them to another operating system - people can just use Linux.

To answer your specific questions (be aware that much of this answer is my opinion - my wife will be happy to tell you, frequently and without much prompting, that I've been wrong before -:)

1. Does anyone here have experiences they would like to share using SANOS?
     No, I'd never actually heard of it until this question popped up on SO (which should go a little towards justifying my statements that it will likely never be important in the grand scheme of things).

2. What is the most expedient way to attract others who would like to contribute time to make it better?
     This is a good start since you've made a large community of geeks aware of it. I'd shop it around some more, you need to get the size of that cadre up.

3. In what ways is it better than either the kernel used in BSD and/or Linux?
     No idea. I have zero interest in this project since I'm more concerned with running a business and spending time with the family than donating my time for free to a project that I think will go nowhere (I'm not trying to be obnoxious or condescending there, that's my actual belief). Given it's minimalist approach, it's possible that it could be faster.

4. Is there merit in porting Ruby to this platform?
     Other than as an exercise in porting, probably not. The Ruby crowd are not going to swarm across to SANOS just because you make it available. They have to earn a living as well.

paxdiablo
+2  A: 

I have to agree with Pax, but I'll take it a step further. If you're really intersted in building a JAVA OS, why not just do it on top of the Linux kernel as a new linux distribution. Start with the Kernel, add just enough user level software to get Java up and running and then start up a java "shell". This way you get all the benefits of Java that you desire without the pain and agony of reinventing the OS wheel. This is exactly what I do with my minicom example in my Embedded Linux book written in 2001. Most of the example is available through Google Books (http://books.google.com/books?id=MFjhhiIMMHIC&printsec=frontcover).

JohnnyLambada
If you decide to go this route, also check out http://www.linuxfromscratch.org/. It's a good overview of Linux internals that allows you to build a stripped Linux distro and has a version for heightened security as well.
Kelden Cowan
A: 

i need document about process managment and memory management sanos? please help me

pari