Can you fork your own distro?
 

I read and see that many people have created custom distros off of mainstream distros (Slackware, Debian, Mandriva, you name it!)How hard a task is it to pull this off, however? And, I know someone may ask, why another distro!! But, to be honest, there are some things I like in one distro and that I like in another distro. Therefore, here is the idea for a another distro! To top it off, I took a computer programming course at a technical school, but not at college. At a college, they may teach you more about operating system design and compiler design. This the knowledge I am still trying to learn.
In a nutshell, is this feasible? To give you an idea, what user would want a distro for example, that has the ease of use, such as PCLinuxOS that wants the speed of distro such as Archlinux or Gentoo, with the stability of Debian, and the flexibility of Slackware? Any thought, rants, or comments would be highly appreciated.

P.S. I watch and read of distros that have one main person, such of as PCLinuxOS, or SimplyMepis. I'm willing to collaborate out with more, however.

I'm sure all of us. But with a couple of hundred distros out there already - some involving quite large teams - don't you think it's a bit of an illusion to do on your own what others have been trying to do for many years now? It seems to me that you can't have your cake and eat it. If you want easy, you'll need to sacrifice speed and flexibility. If you want speed, you'll need to start stripping parts out, making things less easy and potentially unstable in the hands of the average user. There is nothing as flexible as Linux from Scratch, but man, do you expect the average reader to wade through all the documentation?

Strictly speaking, you fork a particular project.
However, you can create your own distro and it is as simple as rebranding anothers distro (see the RHEL clones) and as complex as programming your own interfaces.

Basically, if you hand it out as your own collection, it's a distro.

Smittix anyone?

Ease of use: this is one of those bugbears. There is no such thing... no matter what you pick, someone will find it hard. Making it easy for "most people" usually just means "most people you know". But that's fine... just make specific claims: say what features make it easy to use.

You also have to consider licensing.

You are probably right. Going from a language from "Basic / Pascal" to a language that is more lower level are two different things. If I were to create a new distro, I think my userbase would be people who were frustrated with other distros, like me. I'm not knocking any distros in particular. I believe each has it's own stength and weakness. I am simply pointing out that it would great to pull the features out of distros I like and integrate them into one system. I realize this would take probably many years. I look at distros, such as SimplyMepis and see how far out of the mainstream distro that they have become. And, you really must give them credit, because think of all the hardware that they have support out there! I think, for example, the "CNR" concept in Linspire is a great concept, because it gives you a greater pull of packages.

I've also heard of the philosophy: "I just want it to work!". I agree with this as well, but if everybody agreed with only that one philosophy, why would be there be other distros out there with more features? I think there have been people with great intentions with distros. Those intentions are tailored to their particular needs. The implementation of that intention, is a different story.

The philosophy I most agree with over the years, has been simplicity, flexibility, stability. Simplicity to me means not just a nice interface, but a simple, overall experience. Flexibility means I can customize the way I want it. And, finally, stability means not only that it won't "crash", but everything just works good together.

I guess this would fall into the "Intermediate" user, because the person would have to have a little experience at it.

Btw, how hard is it to create your own distro with LFS? I heard another distro mentioned somewhere was Crux as a plain distro, similar to LFS. Any thoughts?

LFS is a good place to gain an understanding of the nuts and bolts of GNU/Linux. However, it is quite difficult to create the kind of distro you are talking about from it.

Give it a go anyway.

You'll discover why most people start from a distro that gets it "almost right".

You should also begin your research by listing what qualities you want as specifically as you can - then look through, and actually try out, a great many distros (50-ish) and see how they address the qualities you want to emphasize. What trade-offs have been made.

This will also give you an idea of the skillbase you need. It may be a matter of assembling packages under an existing framework, or you may need to write some interfaces. This will tell you.

Attracting people to your project is tougher. New distros have more room at the top, so that's a plus. However, you still need to be able to explain why your distro is more worth someone's time than an established one that is almost like it.

FWIS: after this exercize, you will have acquired a powerful skillbase which will be a benefit to you and the community, even if your distro is a flop. So, I encourage you to steam on.

Good luck.

Some of the reasons you have mentioned is exactly what made me build my own live cd's/dvd. Most of them have a particular purpose, I made a version of slackware 11 that used AES loop encryption to completly encrypt the live cd when you installled it to hard drive. It then created the decrypting boot cd you needed to have in to boot up, otherwise you could not decrypt any any of the partitions.

also finished last year, a live email server that runs Qmail, clamav, spamassassin, squirrelmail, qmailadmin and vqadmin, plus a few other nice things. This is based on slackware 11 too, and currently working on a slackware 12 version, which is not too far away from completion. As far as I know there is no current live cd/dvd mail server getting around. So hopefully it will be unique and I can give something back to the community that has help me so much in the past

Well, for that matter, the fedora project has attempted to make custom distros easier.

It wouldn't be hard with LFS at all, you would need to do some package management, but after that it's pretty straight forward, it's just a matter of how much time you have and what you want to include.

I personally use pkgtools as my package manager and some custom slackbuilds, there not compatible with slackwares, but they do the same thing.

Well, difficult to say.

It terms of "having a Linux up and running X included and some kind of "of course I also have to create my own init stuff" - not that difficult. It's a really educational project - I had much fun and learned _a lot_.

First of all, I learned it's absolutely NOT simple to create something smooth and easy to install with nice and useful options configured - even just for my own very specific needs. ;) After a few weeks, I was really tired of "Oh no, I forgot XY!" - because it's the nasty little things which make a distribution pleasant to use and to work with.

I personally would do it again - just for the fun and sake of of it - but just as a personal project - there're really enough distributions out there. Maybe something small - similar to the Linux router projects or something embedded - stuff like that.

smitty654 are you serous about a fork?
 

I am get more pissed off about how gentoo is being every day. Sooner or later I will fork gentoo or help someone if they beat me to it.

Try Linux From Scratch :)

Well, that sounds to me as if you would like the best of all distros and combine it. That might not be so easy, because "ease of use", "speed" and "stability" are not things you can (de)select in a menu and then you have it. They depend on lots of factors that are sometimes extremely complicated. And sometimes they depend on your computer, hardware and your preferences. A distro might be fast for you but someone who runs other applications on it might find it extremely slow.

LFS is a place to start, as my previous posters all said. Lots of trying with LFS or even kernel hacking might get you where you want to go. But don't forget that it will be frustrating from time to time, but it will be worth the effort. And set yourself realistic goals, e.g. do some small projects first before you attempt to attack you personal distribution that is stable, fast and easy to use all at the same time. Experiment and enjoy!