Re: (long) Re: Why would I want an LFS system?
[cc me please]
On Tue, Jan 29, 2002 at 05:01:40PM +0100, Csan (Ja'nos Hola'nyi) wrote:
> Still, i feel like on the URL I included from linuxfromscratch.org, your text
> could be interpreted in an ambigous way.
I've come to realize that too (I've been thinking and re-reading the intro
page after I wrote my email) and as such I'll remove the specific
references to distribution names and just put a generic term like "regular
distributions" (as opposed to LFS that's not really considered regular in
anybody's mind). I see how people can get offended by me calling Debian or
Redhat by name. It was only meant to give an example of what I consider a
regular distribution, but I'll just remove it and leave it up to the reader
to decide for him/herself.
> the readers draw - btw false - parallel assumptions against the concurrent projects.
Agreed.
> As also somebody else on this list mentioned, the same applies to ./configure &
> make install
> Doesn't it?
If you just run that, then yes there's nothing to gain (well other than
seeing a compile job at work and you'd get familiar wiht the output and
grow comfortable compiling things in the long run, but it doesn't teach you
more). LFS isn't about teaching how to read and understand source code (at
least it's not teh book's aim. But people can alwyas choose to do this
themselves).
So what do you learn from compiling that a pre-compiled binary doesn't give
(at least that's what I think. If somebody doesn't agree, well we live in a
free world ;o)? If you run a simple thing like "configure --help" you can
get a feel of what you can modify in a package. Then people really should
read the README and INSTALL files that come with packages to see if they
want alternative installations. But people will definitely learn what a
package contains and why it's important to have it. Often a package is
important just because "that other package depends on this one". But
there's often more to it than simply resolving dependency requirements.
The LFS-Book itself doesn't go in detail on what exactly the importance of
a package is. I could write a whole book on just one package (and that
multiplied by about 56 becomes a lot of work to maintain). So if people
don't take time themselves to investigate things, try out other
installation methods, then they might as well use a normal distribution and
save themselves a lot of time. Ther's nothing to gain then.
> This list is debian-curiosa, and as such, I felt like that kind of little pun
> could be tolerated (the 'debianfromscratch' part) ;)
It's actually not a bad idea I think. There are people on the LFS
mailinglist I've seen who don't build LFS systems. They use the book as a
guideline to build <insert favorite distro> from scratch (i've seen it done
to Mandrake and Slackware).
> E.g. If you spread your files all over your system following your own rules than
> you will have real hard times finding somebody to help in case of need.
True. Consider this:
If you have enough Linux experience to be able to spread files around all
over the place and breaking the standard conventions (like the FHS), then
you most likely know exactly what you are doing and are not going to need
any help from people that need to login to your system and do stuff.
If you don't konw what you're doing, follow the insatll. instructions from
the book more closely and you get a filesystem hierarchy that's pretty much
100% FHS compliant (it's not 100% yet, we're still working on fixing a few
remaining things). Every distro should be FHS compliant to a certain degree
so Debian people shouldn't have a hard time finding files on an LFS system.
> I hope you don't feel offended like I (also) do when reading that page.
I'm not offended. I actually never realized the intro page could be taken
that offensively so I'll make sure it's changed by the end of today.
--
Gerard Beekmans
www.linuxfromscratch.org
-*- If Linux doesn't have the solution, you have the wrong problem -*-
Reply to: