At 07:29 PM 3/21/00 +0200, Al <[EMAIL PROTECTED]> wrote:
>Will the hardware crash or the software crash?

Yes. Just the other day, I was working on a laptop with a dial-up TCP/IP
connection. Suddenly, the system froze. There was no message. There was no
explanation. It just froze. I restarted the computer and was unable to get
re-connected. It took another person to point out that I had accidentally
unplugged the telephone cable while leaning on it with my elbow. I has
broken off the RJ-11 connector.

I want a new OS. I want a robust system that (1) does not freeze up when
the modem looses a connection, and (2) offers some explanation for why the
PPP connection went down. It should display a message such as "check your
telephone or modem cable", shouldn't it?

Both inside and outside the laboratory, there are a lot of things that can
and do go wrong. This is why modern languages use exception-style programming.

When you "crash" a car, part of it might be temporarily out of service.
When you "crash" a party, a few people attend without invitation. Something
has "crashed" when it no longer performs all of its intended function.
Every exception that's thrown is a crash in miniature.

The "crash" issue is different than the "automatic crash recovery" issue. A
keyboard "crashes" when it's unexpectedly unplugged. Typically, a non-USB
keyboard does not have automatic crash recovery. Plug it back in and it
still doesn't work.

A computer needs power. It "crashes" when its power cable is unexpectedly
unplugged.

A network interface card "crashes" when its network cable is unexpectedly
unplugged.

A mouse "crashes" when it's unexpectedly unplugged.

A modem or fax machine "crashes" when its telephone or serial cable or USB
cable is unexpectedly unplugged.

A printer or scanner "crashes" when its parallel cable or USB cable is
unexpectedly unplugged.

A classic virtual machine "crashes" when an Error is thrown. Look at the
following code:

  public void example() {
    throw new Error();
  }

A kernel "crashes" when it unexpectedly runs out of memory and has no
virtual memory manager. A kernel "crashes" when virtual memory unexpectedly
runs out of free disk space. A kernel "crashes" when a swap particition
unexpectedly becomes corrupt or lost by a physical defect in a hard drive.

When application software is defective, it can "crash" an application. When
system software is defective, it can "crash" a kernel.

I am sure that they crash because both hardware and software sometimes do
not perform their intended function. All these things have crashed in front
of me. And for most, it has happened more than once.


_______________________________________________
Kernel maillist  -  [EMAIL PROTECTED]
http://jos.org/mailman/listinfo/kernel

Reply via email to