On Mar 3, 3:18*pm, (Doug Miller) wrote:
In article , wrote:





On Mar 3, 10:54=A0am, (Doug Miller) wrote:
In article =
..com, wrote:

On Mar 2, 9:01=3DA0pm, (Doug Miller) wrote:
In article , (D=
on =3D
Klipstein) wrote:
In , Doug Miller wrote:
The thing that really stood out to me was the statement by Toyota's
president that they're going to look into programming a brake overri=
de
for the throttle.

I have only one question: WHY IN GOD'S NAME WAS THAT NOT THERE FROM =
THE
BEGINNING?

=3DA0*Programming* a throttle override by the brake? =3DA0As in rely=
ing on =3D
lack
of electronic malfunction in order to have the brake reliably apply a=
n
override onto the throttle?

Since the override becomes necessary only in the event of a throttle
malfunction, for the override to not work would require a second malfu=
nction.
Clearly two simultaneous malfunctions are *far* less likely than any s=
ingle
malfunction.

That's obviously totally false.

No, in fact, that's an elementary principle of probability theory: any tw=
o
events in combination are less likely to occur than either one of them al=
one.- Hide quoted text -

- Show quoted text -

That's false too. * The probability of two events occuring in
combination is only less IF THE TWO ARE INDEPENDENT. * You are arguing
that it's perfectly fine to have the same computer that is running the
throttle to also be the safety override and to disengage the throttle
if the brakes are applied.

I said no such thing.



*Running on the same computer, those two
events are no longer independent. *

Indeed, that is so. But "running on the same computer" is *your* idea, not
mine.



Here is the first post from Don and your reply:

Don:
*Programming* a throttle override by the brake? As in relying on lack
of electronic malfunction in order to have the brake reliably apply an
override onto the throttle?


Doug:
Since the override becomes necessary only in the event of a throttle
malfunction, for the override to not work would require a second
malfunction.
Clearly two simultaneous malfunctions are *far* less likely than any
single
malfunction.


In the context of the discussion here, it seems very reasonable that
Don's logical meaning was that if you program the brake safety overide
on the same computer that is controlling the throttle, then you're
potentially exposed to the same fault. A computer malfunction that
caused full throttle could also result in the same computer not being
able to perform the brake safety function.

You could have just said, it's OK if it's programmed into a SEPERATE
independent computer. That would have added clarity instead of your
reply, which only made it more confusing. And your statement as made
is WRONG anyway, because the requirement for two simulataneous
malfunctions is only true if the program resides in a SEPERATE
computer. That qualification you never made. You seem to expect
everyone else to spell out all the conditions and qualifiers yet you
yourself leave things vague or confusing and think it's just fine.