The Doctor What wrote:
> From a users prospective it doesn't make much sense for all these
> options. I'm using a hypothetical user, who is savvy enough to
> build a kernel. Barely.
It's quite difficult to predict what a user knows - e.g. most people
who buy some packaged "plug and play" system don't have much of an
idea of what is inside. However, you can be sure they know _exactly_
name and number of that new 3D-accelerated video card they're buying
the week after ;-)
What would be generally useful is a way to use the current system
configuration as a default. This eliminates the worst mistakes (e.g.
the (in)famous "VFS: Unable to mount root fs" panic). The rest of
the confusing details could be dealt with by properly structuring
the menus, and maybe a global search option.
> Given that info (plus maybe a little more), it should be possible to
> build a kernel, with everything modular.
I think many distributions are moving in this direction. However,
there are many modules that you probably don't even want to try.
E.g. it sends shivers down my spine when the Elektor I2C driver
happily announces to have found its chip ... on my PCMCIA Ethernet
card :-( (Why I have this driver in some of my kernels is another
story ;-) Things are getting better with the extinction of ISA,
though.
> A simple userspace external tool could be used to link modules
> together into the kernel for cases that need that.
This is actually something we've discussed on linux-kernel for many
years (5 ? Even more ?). It never got written, though. Fortunately,
modules and initrd now cover most of the cases where this would be
useful.
So, who's going to write that hinv --config-syntax or an
"intelligent" version of that /proc/config that's making regular
appearances in linux-kernel, a bit in the style of the monster of
Loch Ness ;-)
- Werner
-- _________________________________________________________________________ / Werner Almesberger, ICA, EPFL, CH werner.almesberger@ica.epfl.ch / /_IN_N_032__Tel_+41_21_693_6621__Fax_+41_21_693_6610_____________________/- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Mar 15 2000 - 21:00:13 EST