OSDev.org

I think it's much easier to make a good, stable, secure server OS than making a good, stable, secure desktop OS. There is a lot of overlap, and you can cut a lot of stuff away in the server project. My Unix core is getting pretty far, it can be a working server OS within a year if I add networking.

sortie wrote:I think it's much easier to make a good, stable, secure server OS than making a good, stable, secure desktop OS. There is a lot of overlap, and you can cut a lot of stuff away in the server project. My Unix core is getting pretty far, it can be a working server OS within a year if I add networking.

Seconded. I, too, have no interest whatsoever in GUIs, and am specifically working on building a headless system. That's, indeed, where an otherwise hobbyist endeavor hits the road to actual production: nobody cares how a useful network service (or these days, cloud service) is implemented under the hood. So long as it does the job and can run as a virtual machine under the likes of Xen or KVM, it can certainly be implemented using a non-mainstream operating system. This as contrasted to the somewhat larger challenge of getting someone to swap out their day-to-day desktop operating system.

sortie wrote: A useful measurement is whether someone is self-hosting, and the various degrees of it:

• Not self-hosting.
• Able to compile small programs under your OS.
• Able to build your OS under itself. (Self-building)
• Able to build your OS under itself on real hardware, reboot, and the new version boots and is capable of the same. (Self-hosting)
• Self-hosting, and able to also build all third party software ports under your OS.
• You actually prefer developing under your OS rather than cross-developing. (Actually self-hosting).

I can call myself 9.5/10 using this scale . Because I am "Self-hosting, and able to also build all third party software ports under my OS" and sometimes, I also prefer to develop under my own OS.
In my opinion, self-hosting scale is very wrong (as in reality, my power level is much lower).

I guess that, as always, whether it is a good metric depends on your goals.

I give myself 1/10. I don't know what I don't know, therefore who am to judge that I know?

I give myself NaN/10.0. \o/

(I've already implemented large portions of the user space but the kernel is way behind )

5/10, My kernel is self-hosting but i havent tried building anything other than simple-ish programs on it, most complicated build on it was klange's nyancat. Besides that it supports reading from ext2, has a simple-ish gui that also works on linux without modification and uses bash (with full job control) as its default shell, combined with busybox for the basic utils

MessiahAndrw wrote:I don't know what I don't know

You are right. I do know that there are a plenty of things that I am not familiar with but I know enough to say that they exist. Then there is an endless pool of things that I am not even aware of. When I learn something, it usually goes from "unknown to me" to "I know it exists" to "I learned it". Let them be steps 1, 2, and 3. The main problem is throughput, i.e. moving from 1 to 2 is a lot easier than moving from 2 to 3. It means that things are piling up at step 2.

In practical terms this means that I will be at the lowest level more and more firmly.

I think its really hard to make an OS that survives contact with users.

Its (relatively) easy to make an OS that works well when you're developing it, and seems promising.

As soon as real users get involved, and as soon as there is any kind of community around it, then bugs start to get found and the fixes for the bugs introduce compatibility problems for your existing user base and you start having to take pragmatic hacks to fix problems without breaking things etc.

Indeed, not breaking interfaces makes things much harder. It's only reasonably doable when you have mature code.

In tyndur we tried to keep binary compatibility between all 0.2.x versions, and much to my surprise, programs compiled for 0.2.0 actually still did run in the latest version (which was almost exactly a year later). But there are so many fundamental changes that need to be made that there would have been no way to keep these programs running in 0.3.x.

MessiahAndrw wrote:I give myself 1/10. I don't know what I don't know, therefore who am to judge that I know?

Same answer I was going to give.

All of us are in the situation of not knowing what we don't know. The difference is that some know that they don't know what they don't know, whereas others don't. The former will give themselves 1/10, the latter 9/10.

I've already laid the ground rules for the 1 to 10 scale. If you can write a working bootloader, you're not a 1. If you don't have a working operating system, or if you are confused by ACPI, PCI, USB, or APIC, you're not a 10. Most people on this site are neither a 1 or a 10.

Let me put it another way. If I have ever answered a question that you posted (correctly), you are between 1 and 5. If you have answered one of my questions (correctly), you are between 5 and 10.

Brendan wrote:On a scale from 1 to 10; the first group aren't OS developers (they're only "OS implementers") and they should be limited to scores from 1 to 5.

Oh dear, what about all those Linux/Unix clones?

Hi,

mathematician wrote:

Brendan wrote:On a scale from 1 to 10; the first group aren't OS developers (they're only "OS implementers") and they should be limited to scores from 1 to 5.

Oh dear, what about all those Linux/Unix clones?

That comment was intended as a joke, however...

Re-implementing an existing design from 40+ years ago avoids a whole pile of decisions (and research, and risk of unforeseen consequences), and then porting an existing user-space avoids a whole pile of work (and decisions, and research, and risk of unforeseen consequences).

It's like comparing an artist that started with something like this (where you know what the end product will look like before they even begin):



..to an artist that started with something like this (where the end product depends purely on the creator's imagination and skill):





Cheers,

Brendan