LiveCD or No LiveCD?
gerard at linuxfromscratch.org
Mon Feb 25 17:39:25 PST 2008
Everybody brought up valid points, as usual. Here's my two cents worth.
Instead of placing myself in the shoes of an average user, let me come
at this from a personal point of view - a nine year old LFS old timer.
Over the years I have found the LiveCD helpful.
If I'm going to install LFS on a machine, either on a real machine or in
a virtual machine environment (such as vmware), I start out with a blank
harddrive or a drive that has a non-Linux system installed already.
Whatever the starting point, the fact is that in such cases I don't want
to have to install a Linux system just so I can install LFS on the same
machine. That way I waste partition space. Maybe the space can be
repurposed later on (as a /home partition when all is done for example
or a /var partition - pick something) but maybe it's not practical to do
so after the fact for reasons I haven't thought of yet.
I'm sure some of you will argue the fact that with today's hard drive
sizes, it's becoming more and more a non-issue. If you have a 200 GB
drive, wasting 5 or 10 GB is not a problem. The real issue is just the
principle. To a lot of people this is the more important issue at hand.
I'm sometimes a bit of a purist so I fall into that category as well.
It's still a very valid point that must be taken into account somehow.
If you setup a temporary partition just so you can bootstrap LFS off of
that, you need to be able to format that partition at the end of the LFS
book and use it for something within LFS. If not, the process is flawed
and needs to be fixed. LiveCD is one fix. Maybe other fixes exist.
Resizing a partition or merging the two isn't always a feasible option.
To me personally the best solution is a LiveCD of some sort. In the end
it matters very little if that is an LFS based system or another
project's. As long as it has a working compiler + all the other tools,
I'll be just fine.
An LFS LiveCD would be a perfect solution as it has all the tools
properly installed. It even comes with the base packages so you don't
need to re-download them all the time.
You still have the educational value of building an LFS system, although
I'll admit using a pre-installed system may be easier to start from. In
that case, you can still access your email in the normal fashion.
Seeing an LFS system isn't installed anymore in a manner of a few hours,
and the fact you need part of BLFS to be fully functional as a user,
which in itself can take days, a LiveCD isn't always practical if your
starting system has Linux installed already and has your personal
profile setup so you can check email, do work and browse the web while
you install LFS in the background.
Taking all those arguments into account, the LiveCD has a place. But the
caveat is that it needs to be actively maintained as part of the core
LFs project, not as a sub-project. That way we can make sure every LFS
book has a matching LiveCD that's up-to-date and works on most hardware.
We may not have the resources (yet) to provide a LiveCD system that
works on every platform, if it works on a lot of platforms then we are
simply providing a very useful feature for a large part of our
community. That's good enough of a reason to continue the endeavour.
On a personal note, I haven't been able to use the CD for a long time
because it won't recognize most SATA controllers that my systems use
(regular Intel ones you find on motherboards nowadays) and thus has
become useless to me for any modern system that I use. I don't know if
the latest version works, but this will be an ongoing problem.
We'll need to decide on a series of available computers that we decide
to be generic systems/chipsets so we can at least try and support the
majority of systems.
Just some food for thought.
More information about the lfs-dev