On Sun, May 28, 2000 at 05:37:55PM -0600, Roberto Mello wrote:
>       What about convenience on booting Linux ? You didn't answer my question
> where I asked if I would have to use a boot disk to boot Linux if I
> didn't install Lilo in the MBR.

I answered that question before you asked it, in my first message.

However, for your benefit I will explain the basics of how a computer
system boots.  After the power-on self-test, the computer's BIOS routines
check the CMOS settings to see what device it should boot from.  We'll
ignore uncommon events like booting from a floppy or CD, and say that
the computer is booting normally, from the first hard drive.  The BIOS
routines pass control to the MASTER BOOT RECORD of the first hard drive.
The master boot record is an area on the drive before any of the partitions.
You can't see it, it's not part of any filesystem, but it's there.  The
MBR (hopefully) contains code for booting the system.  Also at the front
of the drive, also before any partitions, is the PARTITION TABLE.  It's
simply a list of partition boundries.  The partition table also contains
an ACTIVE (also known as BOOTABLE) bit which can be turned on or off for
each partition, although for the system to work properly only one partition
can be marked as active.  The (default) code in the Master Boot Record
checks the partition table to see which partition is marked as active,
and then passed control to the BOOT SECTOR of that partition.  This is
where LILO should be, if you were smart enough not to install it to the
MBR.

>       Is this a volatility of the MBR or an assumption by Windows that the
> user is only running it ?? I don't think it's a volatity of the MBR, the
> MBR is doing nothing, Windows is the one who wrote on it.

When I say "volitility" of the MBR, I mean that many programs overwrite it.
Many OS's overwrite it.  Many virus overwrite it.  No, Windows doesn't make
the assumption that the user is running only it.  Windows makes the
CORRECT assumption that *NO OPERATING SYSTEM* should be putting anything
other than the normal MBR code in the MBR.  It makes me sick to say this,
but in this respect, Windows is actually more sensible, friendly, and
standards-compliant than most Linux distros, by not anally rapeing the MBR.

>       What if I want to dual boot between windows and Linux ? Will I need to
> mark my active partion for everytime I want to change the OS I am
> running ? If it is this way, than it IS cumbersome, and there's an
> advantage of putting LILO in the mbr.

No.  That would be silly.  If it were that way, would I be saying there
was no advantage to putting LILO in the MBR?  You must not think very
highly of my sensibilities.  But no matter.  No... LILO on your Linux
partition works exactly the same way as LILO in the MBR from the user's
perspective.  The only difference is that by putting LILO in the MBR,
you lose a certain amount of flexibility.

It boils down to this:  with LILO in the MBR, the only way to *NOT*
boot through LILO is to uninstall LILO, format the MBR, or boot from
some other boot device.  That's absurd foolishness.  If LILO is installed
to the MBR, you *ALWAYS* have to boot through LILO.  Foo.

>       I don't doubt it breaks, it just never broke with me in two years. When
> it breaks, boot Linux or windows from a floppy and rewrite it. Not too
> hard I think.

I debate zealous LILO-in-MBR fanatics on a regular basis.  There's no
logical basis for their beliefs, and some of their arguments really
bother me.  For example, one of the most common arguments I get from
them is "If the MBR gets wiped out, it's not hard to reinstall LILO
to it."  However, tell me which of these is "easier" to do:

1.  Something.
2.  Nothing.

QED.

-- 
Craig McPherson
[EMAIL PROTECTED]

The world's funniest joke:
"Memes are a hoax.  Tell all your friends!"


_______________________________________________
Stormlinux-users-list mailing list
[EMAIL PROTECTED]
http://www.stormix.com/community/lists/listinfo/stormlinux-users-list

Reply via email to