Anthony E. Greene <[EMAIL PROTECTED]> wrote:
> Douglas Alan wrote:
> > No it wouldn't. It is never reasonable to destroy large amounts of
> > data without being quite sure that that is what the user wants.
> If that were true, then 'rm -i' would be default behavior, and the
> '-f' option would not exist.
"rm -i" would never be the default in a good operating system because
endlessly nagging the user is completely the wrong approach to making
good resilient software. On the other hand, no modern operating system
would be designed so that the standard file delete command permanently
deletes files without an opportunity to undo the delete. We are stuck
with this in Unix for historical reasons. Regarding "-f", it should
definitely be a more verbose option. The reason that it isn't, is again
historical.
> Clearly, there are situations when the user is expected to know what
> they're doing. I think that creating an automated Linux install config
> is one of those situations. You don't.
Please don't completely mischaracterize everything I have said. I
haven't said anything like the the user shouldn't be expected to know
what he is doing.
The people who operate a nuclear power plant should know what they are
doing, right? Does that mean that one wrong flicked switch should cause
the plant to instantly melt down?
> Kickstart is a non-interactive environment, isn't it? How is it
> supposed to interact with you to confirm your instructions?
It could start by not zeroing partitions on disk drives uninvolved in
the OS installation, since there is no reason for it to do that.
Other improvements might be for it to put up a splash screen at the very
beginning of the process, detailing exactly what the installer is going
to do, which disk drives it is going to muck with, and which partitions
it is going to destory, and then ask the user to type "confirm" or
somesuch. But I'm happy to start with small improvements first, like
merely not zeroing partition tables on disk drives uninvolved in the OS
installation.
>>> I have never done this kind of automated installation *precisely*
>>> because I did not want the installation routine to make decisions
>>> for me.
> > Now you are making no sense at all. First of all, Kickstart does
> > nothing that the interactive Red Hat installer doesn't do. The
> > exact same issue comes up in the interactive installer. Are you
> > saying that you don't use the Red Hat installer at all?
> No. I was talking about what I want. Not what I put with. I take
> actions to make the situation conform to my wants, limited by the
> amount of time/effort I'm willing to invest. In this case, I'm willing
> to invest the time/effort to run an interactive install so that I can
> hae more control over what's happening.
The Red Hat interactive installer gives you no more or no less control
over what's happening than the Kickstart installer. Both present you
with almost exactly the same set of options.
> > Furthermore, you say that you won't use Kickstart because it "makes
> > decsisons for you". I stand here saying that it should't make decisions
> > for you. You disagree with me and say that it *should* make those
> > decisions for you,
> I said no such thing. What I said was that it's action was reasonable,
> given the assumptions it works under.
Well, then, you are wrong on that assertion. The behavior that I am
complaing about in the Red Hat installer -- that it zeroes the partition
tables of disk drives uninvolved in the installation -- is wrong
behavior. It is indefensible and should be fixed.
> Even if your characterizations of my comments were correct, your
> characterizations assert an inconsistency where none exists. Here are
> the points you attribute to me:
> 1. I don't use kickstart because it makes decisions for you.
> 2. Kickstart should make decisions for you.
> Those two statements are not mutually exclusive. They make just as
> much sense as:
> 1. I don't buy cars with automatic transmission because they make
> shifting decisions for you.
> 2. Automatic transmissions are supposed to make shifting decisions for
> you.
You forgot to add your third claim:
3. And because they make shifting decisions for you, don't whine
when you get on the Autobahn and the automatic transmission
decides to floor it for you until you're going 180 mph.
Just because an automatic transmission makes shifting decisions for you,
does not mean that those shifting decisions are beyond the scope of
engineering criticism. In fact, they most surely are within such
scope.
|>oug
--
redhat-list mailing list
unsubscribe mailto:[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
