Information Technology Reference
In-Depth Information
s
1
)]
:ab
2
[
up
(
s
2
)
broken
]
in conjunction with these three steady ones:
:ab
1
[
(
light
up
broken
_
malfunc
1
_
wiring-problem
:
light
ab
1
broken
_
malfunc
1
_
wiring-problem
ab
2
malfunc
2
_
wiring-problem
Let the underlying influence information be
I
=
f
(
up
(
s
1
)
;
light
)
;
(
broken
;
light
)
;
(
malfunc
1
;
light
)
;
(
;
)
;
(
(
s
2
)
;
)
;
(
malfunc
2
;
)
;
wiring-problem
light
up
broken
broken
(
;
)
;
(
; ab
1
)
;
(
malfunc
1
; ab
1
)
;
wiring-problem
broken
broken
(
; ab
1
)
;
(
malfunc
2
; ab
2
)
;
(
; ab
2
)
g
wiring-problem
wiring-problem
This determines the following causal relationships, all of which except for the
rst three are steady:
(
s
1
)
causes
if
:ab
1
up
light
:
up
(
s
1
)
causes
:
light
if
:ab
1
up
(
s
2
)
causes
broken
if
:ab
2
causes
:
light
if
>
broken
causes
:
if
>
malfunc
1
light
causes
:
light
if
>
wiring-problem
causes
ab
1
if
>
broken
causes
ab
1
if
>
malfunc
1
causes
ab
1
if
>
wiring-problem
:
causes
:ab
1
if
:
malfunc
1
^:
broken
wiring-problem
:
malfunc
1
causes
:ab
1
if
:
broken
^:
wiring-problem
:
causes
:ab
1
if
:
^:
wiring-problem
broken
malfunc
1
causes
ab
2
if
>
malfunc
2
wiring-problem
causes
ab
2
if
>
:
malfunc
2
causes
:ab
2
if
:
wiring-problem
:
wiring-problem
causes
:ab
2
if
:
malfunc
2
Consider, now, a situation where we only know that both switches are open
(as depicted in Fig. 4.1). What would be the predicted outcome of rst tog-
gling
s
2
, followed by
s
1
? Since nothing hints at a malfunctioning battery or
bad wiring, we expect that
up
(
s
2
)
broken
be true and, hence,
broken
is
an indirect eect of closing
s
2
. Consequently, closing
s
1
afterwards should
