Separate book? (was: Bootscripts merge?)

Alexander E. Patrakov patrakov at ums.usu.ru
Mon May 29 08:40:44 PDT 2006


Bruce Dubbs wrote:
> On 5/29/06, Alexander E. Patrakov <patrakov at ums.usu.ru> wrote:
> 
>> Simply because the adequate (i.e.: with good educational value and 
>> technical
>> correctness) explanation of the udev topic is way too thick for the 
>> LFS book.
> 
> Why?  BLFS is up to about 1100 pages and LFS is somewhere around 300.
> There is plenty of room for more.

The balance should be kept. LFS is not only a book about bootstrapping a 
toolchain and configuring udev.

>> Also, this in-depth explanation doesn't fit for those who read the 
>> book for the
>> first time.
> 
> But in-depth explanations can be added as appendicies that first time
> readers don't need to read to do a build.  I've always been in favor
> of adding all the bootscripts, and now udev rules, as appendicies to
> make the book "complete."  There is a lot of great information in
> those scripts and files, but I suspect most users just install them
> like any other package and don't look inside.

Appendix does look as a good idea for getting a book that is both educational, 
technically correct and not overwhelming for newbies. Thanks for the suggestion, 
it is indeed an optimal plan for LFS.

However, I am not sure what to do with the following points:

1. Target audience. It is impossible to understand udev and kernel fully without 
reading the source. So, at least for some sections, a working knowledge of C 
must be assumed. AFAIK, this is unacceptable for the LFS book.

2. Other points suggested by Ag: RAID and volume management.

3. Not-so-unusual boot media, such as fakeraid cards commonly found on modern 
motherboards.

Currently, it is impossible to dual-boot LFS and Windows XP from such cards 
without deviating from the book. Reason: the boot process needs the "dmraid" 
program in initramfs, and initramfs is beyond even BLFS, because of its apparent 
complexity, and because in the most typical case (a regular IDE drive) it is not 
needed.

Thus, the detailed and 100% working book becomes essentially non-linear (like 
BLFS). Do you really want to break the tradition and make LFS non-linear? My own 
opinion is that only chapters 4-6 must be linear, but the majority seems to 
disagree.

-- 
Alexander E. Patrakov



More information about the lfs-dev mailing list