Bug#69151: provide advice about partitions to create and their sizes
"MS" == Martin Schulze <joey@finlandia.Infodrom.North.DE> writes:
Joey,
CMC> It would be nice if there were some information about the
CMC> number and size of partitions the user should create (how
CMC> would a new user have any idea that they need at least two
CMC> partitions -- root and swap, and that those partitions
CMC> have to be special types, or have special names?), as well
CMC> as a quick overview of the commands available within pdisk
CMC> (especially how to get help, how to save the partitions
CMC> they've created, and the fact that you can quit pdisk
CMC> *without* disturbing your existing partitions!). It would
CMC> be even better, of course, if the partitioning tool were
CMC> more user-friendly, but the installer should try to make
CMC> up for the weaknesses of the tools where necessary.
MS> I'd like to refuse this request since it only blows up the
MS> boot floppies. It's all written in the installation
MS> manual, isn't it?
Adding a screen of information summarized from the manual ``blows
up the boot floppies''? I can't see how that's possible, unless
you mean ``blows up'' in the sense of ``makes larger'' rather than
``causes to explode''?
And, yes, there is some excellent information about partitioning
in the installation manual, complete with suggestions about the
best arrangement. But assuming that people have actually read the
installation manual before starting an installation is naive.
Especially if some significant number of your users are coming
from another distribution that includes enough information in its
installation program to allow them to do an install without
reading the manual.
In any case, I can't see much in the way of justification for the
space vs. ease-of-use tradeoff to be decided in favor of space.
More and more people do installs from CD-ROM or from the network
as the size of the distribution and the speed of network
connections increase. Crippling your installer to satisfy the
most space-starved installation method doesn't make sense to me.
Perhaps there could be a minimal installer that depends on the
user being familiar with information in the installation manual
and offers only the most basic documentary support, and another
installer that takes advantage of higher bandwidth media to offer
a more information-rich installation experience.
MS> Speaking of pdisk, the powerpc partitioner, documenting it
MS> should be done in the documentation as well.
Again, yes, pdisk should be documented in the installation manual.
But documentation or no, pdisk is extremely non-user-friendly, and
a better solution would be to replace it with another partitioning
tool that is more user-friendly or to add some functionality to
pdisk to make pdisk more user-friendly.
My point in filing these bugs is to try to emphasize that a user's
installation experience is important. Once Debian is installed,
it's a dream to run -- keeping software up-to-date is trivial,
administration is pretty easy because of the way files and
directories are organized, and the system is extremely
well-integrated. But getting to that point can be a nasty
experience, and may well scare people away who would greatly
benefit from running Debian.
Sadly, it's also true that once someone has had a bad experience,
that bad experience will tend to influence their opinions about
future releases, no matter whether they've seen them or not.
Thus, someone who has a bad installation experience now on a home
or experimental machine may at some future point reject Debian as
a solution for other roles because of that bad installation
experience.
I can see the argument for keeping boot floppies mean and lean
(although it's interesting to note that Red Hat manages to fit
everything necessary for a network or CD-ROM install onto one
diskette). But I also think it's important to consider the impact
that a more user-friendly installer could have on Debian's
adoption by end-users.
Pretty X installers with plenty of information about what's
happening and games to play while packages are installed don't add
any value to the installation itself, but they make users happy
because they feel like they feel like they're a part of the
installation process and have something to do while waiting for
the machine to complete some task. Having a text-only ``expert''
installer as an option allows power users to feel a greater sense
of control and fine-tune parameters the average user won't
understand or care about, giving the power users a greater sense
of control.
On the other hand, text installers that ask cryptic questions
without supplying enough context to understand the consequences of
the available choices; assume that the user has either read the
installation manual and planned everything out ahead of time or is
a hacker and knows enough to improvise solutions to any problems
they encounter; and walk the user through procedures that fail on
their machine make the user feel like a victim of the installer.
That's Not A Good Thing (tm), and we can do better.
Finally, I'm not at all convinced that the current installer is
fixable at all. I'm concerned about the next generation
installer, and I think that it's extremely important for
user-experience concerns to play a role in its design.
CMC
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
Behind the counter a boy with a shaven head stared vacantly into space,
a dozen spikes of microsoft protruding from the socket behind his ear.
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
C.M. Connelly c@eskimo.com SHC, DS
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
Reply to: