<s>
A	O
test	B-General_Concept
script	I-General_Concept
in	O
software	O
testing	O
is	O
a	O
set	O
of	O
instructions	O
that	O
will	O
be	O
performed	O
on	O
the	O
system	O
under	O
test	O
to	O
test	O
that	O
the	O
system	O
functions	O
as	O
expected	O
.	O
</s>
<s>
There	O
are	O
various	O
means	O
for	O
executing	O
test	B-General_Concept
scripts	I-General_Concept
.	O
</s>
<s>
Automated	B-Application
testing	I-Application
.	O
</s>
<s>
Test	B-General_Concept
scripts	I-General_Concept
written	O
as	O
a	O
short	O
program	O
can	O
either	O
be	O
written	O
using	O
a	O
special	O
automated	O
functional	O
GUI	O
test	B-Application
tool	I-Application
(	O
such	O
as	O
HP	B-Application
QuickTest	I-Application
Professional	I-Application
,	O
Borland	O
SilkTest	O
,	O
IBM	B-Application
TPNS	I-Application
and	O
Rational	O
Robot	O
)	O
or	O
in	O
a	O
well-known	O
programming	O
language	O
(	O
such	O
as	O
C++	B-Language
,	O
C#	B-Application
,	O
Tcl	B-Operating_System
,	O
Expect	O
,	O
Java	B-Language
,	O
PHP	B-Application
,	O
Perl	B-Language
,	O
Powershell	B-Application
,	O
Python	B-Language
,	O
or	O
Ruby	B-Language
)	O
.	O
</s>
<s>
Automated	B-Application
testing	I-Application
may	O
be	O
executed	O
continuously	O
without	O
the	O
need	O
for	O
human	O
intervention	O
,	O
they	O
are	O
easily	O
repeatable	O
,	O
and	O
often	O
faster	O
.	O
</s>
<s>
Automated	B-Application
tests	I-Application
are	O
useful	O
in	O
situations	O
where	O
the	O
test	O
is	O
to	O
be	O
executed	O
several	O
times	O
,	O
for	O
example	O
as	O
part	O
of	O
regression	O
testing	O
.	O
</s>
<s>
Automated	B-Application
tests	I-Application
can	O
be	O
disadvantageous	O
when	O
poorly	O
written	O
,	O
leading	O
to	O
incorrect	O
testing	O
or	O
broken	O
tests	O
being	O
carried	O
out	O
.	O
</s>
<s>
Automated	B-Application
tests	I-Application
can	O
,	O
like	O
any	O
piece	O
of	O
software	O
,	O
be	O
poorly	O
written	O
or	O
simply	O
break	O
during	O
playback	O
.	O
</s>
<s>
A	O
trained	O
manual	O
tester	O
can	O
notice	O
that	O
the	O
system	O
under	O
test	O
is	O
misbehaving	O
without	O
being	O
prompted	O
or	O
directed	O
;	O
automated	B-Application
tests	I-Application
can	O
only	O
examine	O
what	O
they	O
have	O
been	O
programmed	O
to	O
examine	O
.	O
</s>
<s>
When	O
used	O
in	O
regression	O
testing	O
,	O
manual	O
testers	O
can	O
find	O
new	O
bugs	O
while	O
ensuring	O
that	O
old	O
bugs	O
do	O
not	O
reappear	O
while	O
an	O
automated	B-Application
test	I-Application
can	O
only	O
ensure	O
the	O
latter	O
.	O
</s>
<s>
One	O
should	O
consider	O
the	O
return	O
on	O
investment	O
for	O
automating	O
any	O
given	O
test	B-General_Concept
script	I-General_Concept
,	O
i.e.	O
</s>
