New ideas

[ASHLEY4]

There are usally two reasons for making a new OS,
1. is to learn whats going on underneth your programs.
and
2. to make a new design.

I see a lot of the number 1, but not many of the number 2, Apart from to learn from, why make copys of what is already available.
For instance why have memory protection like ring0, ring3 when this does not stop trojens, virus's and worms infecting our pc's, but stops programmers accessing hardware etc. and slows our applications down.
Take the xbox os as an example, it is only 150k it has no memory protection and will run 90% of all windows applications, with very little modification, and will come with IE and media player 9 in the next version, goes on line and does not seem to be plaguede with virus's ,trojens etc as more and more things are added to the xbox which is basically a x86 pc.
It will turn in to a modern day AmigaOS , right under the nose's of the people that should of been making OS'S like this
YOU

:-[

[frank]

There are usally two reasons for making a new OS,
1. is to learn whats going on underneth your programs.
and
2. to make a new design.

Reason 2 is never a reason, this is neccessary to make something which works. It's not something I always look foreward to.. It's just neccessary if you don't want to write crap.

You are forgetting a few reasons
3. Fun (most important one)
4. Learning about the hardware and Operating Systems.
(I'm learning a lot which I didn't know)
5. Some people also develop oses 'cause they do not like the current OS-es. (such as linux,windows etc)

I see a lot of the number 1, but not many of the number 2, Apart from to learn from, why make copys of what is already available.

You do not make a copy
You make a whole new thing which doesn't exist.
It uses the same hardware yes, but it's totally different.

I suppose it's difficult to understand if you've not done os development yourself.

Take the xbox os as an example, it is only 150k it has no memory protection and will run 90% of all windows applications, with very little modification, and will come with IE and media player 9 in the next version, goes on line and does not seem to be plaguede with virus's ,trojens etc as more and more things are added to the xbox which is basically a x86 pc.
It will turn in to a modern day AmigaOS , right under the nose's of the people that should of been making OS'S like this
YOU

I don't know AmigaOS but I suppose you don't like it very well.

How the OS is just depends on how much work is put in the Operating System and the experience of the coders.
Linux and FreeBSD also started as "pet-os"

[nullify]

For instance why have memory protection like ring0, ring3 when this does not stop trojens, virus's and worms infecting our pc's, but stops programmers accessing hardware etc. and slows our applications down.

You're kind of mixing up the term "protection" here with "security." Memory protection can be used to isolate crashes in individual buggy programs so that the entire system isn't brought down. Usually the operating system implements additional security functionality as well (for example: the UNIX permission bits and user/group access, or more recently ACL's). Bugs in *these* portions of the system are typically what allow viruses, worms, etc. to be effective, not merely the presence (or absence) of memory protection.

P.S. You forgot the all-important idealistic reason to develop an OS: world domination

[Solar]

No, I think ASHLEY4 liked the AmigaOS very much. At their time, they uttered "memory protection is never perfect, and if we can't make it perfect, why bother?"

But, ASHLEY4, the lack of memory protection, and the ability to circumvent the OS API and bang the hardware directly, has hurt the AmigaOS much in its later days. The transition from OS 1.3 to 2.0 was harder than it should have been because of the many HW hitting software; same goes for the transitions OCS -> ECS -> AGA chipset, because things changed and the software wasn't prepared.

And it might have been OK for hobbyists to have a "Guru meditation" rip apart their system and requiring a reboot. But today... no, not having memory protection is not a feature. Add to this that today's CPUs are highly optimized to have memory protection, while in Amiga times a MMU was a special, expensive, add-on chip instead of integral part of the CPU.

As for the rest...

I see a lot of the number 1, but not many of the number 2,

True for many, but not for all. For me, the reason to start my project was to take the lessons learned, and start anew to build a better OS from the ground up. I can promise you, while it's not exactly the ring 0 / ring 3 memory protection, that my system - if I ever shall finish it, which is a different question - will make it much harder for worms and trojans to achieve anything.

Don't judge too quickly.

[Pype.Clicker]

better protection against external attacks is one of the point in Clicker too. I already gathered a few techniques like non-executable data segments (which are absent from current PC-linux implementations, but achievable), cross-checking the call to important system traps (like EXEC() or OPEN()) with the program's relocation list to know if the caller meant to exec() when the program was loaded or if the invokation is the result of some unexpected thing ...

Restricting the view of the FS a program has should help securing programs too ... if your mail client cannot see other files that its configuration and your mail folders, a macro virus or an exploit cannot remove all your MP3s from your hard disk, can it ??

At the 'new design' level, you might give a look at my Ideas page and see that Clicker is not YetAnotherIsNotUnix(tm) but SomethingCompletelyDifferent(tm)...

[darklife]

Remember MS-DOS? lol. I'm learning about OS internals simply because I want to know everything about them that I can get my hands on. I don't know exactly why, but I find OS theory and design very VERY interesting. Maybe I'm just a total nerd, who knows. I?m pondering on whether I want to design an OS or not. At one time, I made a simple real-mode single process OS (Deviator OS), much like CP/M. I had some fun creating it and even using it but I soon abandoned it since I wanted something "more". I think the fact that I'm reading Operating Systems - Design and Implementation, says it all.
It seems hard to design something new because a lot of stuff has been done already. OS's have been around since the 50s. That?s a heck of a lot of years in "computer years"
Once in a while I will think of something that is actually a new idea, and I jot it down quick. I have tons of notepads laying around the house with notes and ideas. I could put them to use if I want. Sometimes new ideas aren't always good. Microsoft has proven this over and over. I'll give them the credit of trying. I hate to say it but I don't really need anything better than what Linux has to offer. Windows xx lacks too much for me, and that's why I don't use it. Linux has everything I need for now. When I need something more, I will make it (application software).