Planning an overall direction for LFS

[taipan]

Robert Daniels wrote:
> I've been thinking about this 'modules' idea,  and also the idea about 
> the automatically generated personalized book, and I came up with the 
> following:
> 
> 1) Bootstrapping a toolchain - Pretty much equivalent to Chapter 5.
> 2) Package Management - Choose your package manager, learn how it works, 
> why they exist.  Maybe a basic scripting tutorial as well.
> 3) Basic CLI - Chapter 6.  Add gpm, choice of text editor.
> 4) Security - How to secure a linux system, cover options like PAM, 
> SELinux, sudo, and whatever else is appropriate.
> 5) CLI Apps - Applications to make a system useful from a command line.  
> Wget, text web browsers, etc.
> 6) Desktops: Probably one of the larger sections.  Cover KDE, Gnome, 
> Xfce, and more basic window managers.  Also include desktop frameworks 
> like Xorg, HAL, and DBus.
> 7) Desktop Apps - Things like office suites, multimedia applications, 
> graphical web browsers, etc.
> 8) Server Software - LAMP stack and whatever else is appropriate for a 
> server, cuz I don't really know.
> 
> A lot of the structure here looks similar to what we already have.  1-3 
> pretty much have to go in order, and I think 4 should also follow 
> immediately after, but the rest is fairly nonlinear and/or optional.
> 

Looks about right to my unskilled eye, but i'd say the 'linearity' of 
modules 1 & 2 would be a bit trickier to implement than it initially 
appears...

If you were to take the 'diy-linux reference-build' as an example, Greg 
caters for the option of package-management during the 'temp-tools 
phase' (or Chapter 5, or module 1, whichever you prefer), which is a 
*must* for any users wishing to take the PM option during Chapter 6.

Furthermore, he provides his own build-scripts, either for illustrative 
purposes or to be used outright, which include 'temp-tools' 
(Chap-5/module-1) in their automation-process.

So sections on scripting/automating & PM ought to at least be linked to 
at an earlier stage in the process than _after_ the bootstrap has been 
completed.

Speaking for myself, i have an almost compulsive need to know _WHY_ i'm 
supposed to do things a certain way, so additional 'pre-reading' would 
be no issue for me, in fact it would be welcomed. But i can envisage 
such an approach driving away the sort of people who have the mindset of 
"yeah, yeah, just tell me where the 'GO' button is so i can build the 
thing!".

Does LFS want to cater to as wide an audience as possible, or would it 
be worthwhile defining a 'target-demographic' at an early stage in the 
planning process?

Feel free to correct any misconceptions i might have voiced - i'm fairly 
thick-skinned. ;)

taipan