Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 26 Jan 2002 15:57:13 -0800
From:      Terry Lambert <tlambert2@mindspring.com>
To:        Anthony Atkielski <anthony@freebie.atkielski.com>
Cc:        freebsd-chat@freebsd.org
Subject:   Re: Why dual boot?
Message-ID:  <3C534259.A20067B2@mindspring.com>
References:  <3C4FBE5C.2AE8C65@mindspring.com> <20020123114658.A514@lpt.ens.fr> <20020123223104.SM01952@there> <3C4FBE5C.2AE8C65@mindspring.com> <4.3.2.7.2.20020124213809.00e6e5d0@localhost> <20020125131659.GB7374@hades.hell.gr> <3C51CD33.4E69B204@mindspring.com> <001b01c1a635$636a4170$0a00000a@atkielski.com> <3C5270E4.BF21F79B@mindspring.com> <011b01c1a659$fb98a670$0a00000a@atkielski.com> <3C52AB34.B8896C8D@mindspring.com> <018c01c1a675$f3dcc1c0$0a00000a@atkielski.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Anthony Atkielski wrote:
> Terry writes:
> > Can you back this statement up?
> 
> Only empirically.

You mean anecdotally.  Empirically would mean you had some
conclusive evidence you could share.

> > Complexity is an emergent property of even
> > incredibly simple-seeming systems.
> 
> But that's not what I said.  I said that the more complex a system is, the
> less stable it will tend to be.

Can you back this statement up?  Have you ever heard the
words "metastable" or "fractal"?


> This has nothing to do with whether or not
> a seemingly simple system is in fact complex.

I said it was an emergent property; this doesn't mean that
"a seemingly simple system is in fact complex"; it means
that complexity will appear, eventually if not immediately.

The key to interacting with complexity is appropriate
abstraction of that complexity along fractal lines.

If we get back to the problem at hand, the partitioning of
disks for the purposes of dual boot, then one of the very
obvious places for abstraction is the fact that there are
two partitioning tools in a self-similar domain, which is
breaking up a disk into smaller areas for particular uses.

So one obvious abstraction that can be had is to have a
single tool that does the partitioning, and to treat the
FreeBSD disk partition table exactly as you treat a DOS
partition table within a DOS extended partition.

This abstracts an entire layer of complexity from the
process, by permitting the same tool UI to be used in the
FreeBSD disklabel domain and the DOS partition table
domain.

Thus, while the system is still just as complex as it was
before, the user experience is less complicated, and the
attack phase of the learning curve is not as prolonged,
since there is one less class of interaction that has to
be learned by the user.


> > Why you are for maintaining the status quo
> > of monumental effort ...
> 
> I'm not.  It makes no difference to me, since I do not build dual-boot
> systems.

You don't act as if it makes no difference to you.  You
act as if you are heavily invested in the status quo.


> > I guess this is OK for a developer ...
> 
> It is okay for a production system or network, too.

Your use of the phrase "production system" carries a lot
of baggage with it; I'm not prepared to grant that the
systems installed from CDROM are "production systems",
unless you are a computer manufacturer who is shipping
FreeBSD preinstalled as part of your system production
line.  If that were the case, I kownt hat you are not
using the FreeBSD install process at all in production,
and may not even be using it in the bootstrap process of
producing a "golden master" disk.  Having had personal
experience with producing a "golden master" CDROM for
embedded systems based on FreeBSD, I can say with
authority that the FreeBSD CDROM production scripts are
not up to the task because it's not possible to use
kernel configs other than "GENERIC", and have sysinstall
correctly install the resulting kernel, since it tries
to copy "kernel.GENERIC", and will not look for the
actual kernel if it is named "kerne.PROD" or anything
other than "kernel.GENERIC".


> Few systems operate in
> isolation these days.

Isolation or non-isolation is irrelevant to the problem, as
far as it has been discussed.  Isolation becomes an issue
when network installation or upgrade are involved, and is
therefore outside the scope of the discussion of CDROM
installs entirely.

-- Terry

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3C534259.A20067B2>