[IanChiles]

Absolutely agree with you. I struggled with wording this for the longest time - and when I talk about OS development, I'm referring much more to the userland development - package managers, core utilities, shells, init systems, and the like, than I'm referring to kernels. Sorry for the confusion.

[blueblob]

I think some of the less explored portions of LFS contain well-documented reasoning for implementing some of these things yourself. For example there is a single solution (out of the many ways to solve this problem) to the package manager dilemma in the hints section of LFS

http://www.linuxfromscratch.org/hints/downloads/files/packag...

which explains reasoning for the package manager and how to implement it yourself (in a much less copy-compile-execute manner than LFS itself).

Implementation is a great way to learn but generally people will survey the literature first to understand why things are implemented the way that they are. Understanding what all package managers have in common will allow you to distill it down to the simplest factor and implement that.

A lot of the core utilities and why they are expected should be covered in the standard (POSIX in this case).

I guess what I am trying to say is that implementation is a useful way to learn but it requires you to know the details whereas understanding the literature first will give you a better idea of what to implement.

Good luck with your project.