My presentation slides for COMPSAC '87. One of the earliest presentation about Combinatorial Testing. K.Tatsumi, S.Watanabe, Y.Takeuchi, and H.Shimokawa, "Conceptual support for test case design," Proc. 11th Intl. Computer Software & Applications Conf.(COMPSAC), Oct. 1987

1. CONCEPTUAL SUPPORT
FOR
TEST CASE DESIGN
Oct. 1987
Keizo Tatsumi
QuaHty Assuranee °ept.
Software Division
Computer Software Development Group
FUJITSU LTD.
Presentaion Slides
IEEE International Computer Software & Applications Conference (COMPSAC '87), Tokyo, Oct. 8, 1987

2. METHODS FOR
BLACK-BOX TESTING
専Equivalence Partitioning
lnput domain of a program are divided into several
equivalent Classes.
1 Boundary- Value Analysis
The edges of equivalent Classes are analyzed.
(e.蛋. Maximum values,Minimum values)
書Cause-Effect Graphing
Test eases are derived from a Boolean graph
linking input and output Conditions.
〔PROBLEMS〕
l Above Methods do not help us in obtaining input
egnditions.
D Ho闇ean闇e Obtain input conditions?
義 Cause-Effect Graphing requires the knowledge and
manPO野er.
Q Ho闇ean闇･e deriye test eases efficiently?
l How ean闇e Share the information and knowledge
for test between testing members?

3. TEST CASE DESIGN
PROCEDURE
l Step-1) TEST CLASSIFICATION
Functions are divided into smaller units.
義 Step-2) TEST FACTOR ANALYSIS
Input eonditions 恒11ed FaetoTS) and their
values (Called States) are identified from
exteTnal specifications.
l Step-3) TEST CASE GENERATION
Test eases are generated by combini鵬the
states of the factors.
義 Step一郎 Definition of Test Results
Expected results are defined for each
generated test eases.

4. TEST FACTOR ANALYSI S
沌Ⅹternal Speeifieations〕
ASCFⅠLEcommand
Command 微 W& 襾2
ASCFⅠLE ｴ 4 ﾔR 66W72ﾖ ﾖR ﾂ
[:::::E'Lf-namef']
凸
ETest Factor Table〕
F a C to 僊 C 認
Command input 4 ﾔV W& 襭
Entrymethod &V蹤 6 aCCeSSーname
r 冤ocation 剿ﾆV誣F
S t a t Terminal VﾆﾆVF B Correctly 磐匁呈ﾗVﾒ(ACNAME) V6貿坊B (lcharacter)
2 綴& F6 Mnemonic 疲 VgB intermediate
e S 剪 4竰 Parenthesis 茶"ﾒﾓv6 & ﾒ 7FW'2
3 Omitted 疲 没③ &V蹤 6 Maximum (8charact- ers)
4 90rmore characters

5. TEST CASE GENERATION
洋est Factor Table〕
F a C to 僊 C 認ﾂ
Command input 4 ﾔV W& 襭
Entrymethod &V蹤 6 aCCeSS name
r 冤ocation 剿ﾆV誣F
S t a t Terminal VﾆﾆVF B Correctly 磐匁儲Vﾒ(ACNAME) V6貿坊B (1character)
2 F6 Mnemonic 疲 VgB intermediate
e S 剪 4竰 Parenthesis 茶"ﾒﾓv6 & ﾒ 7FW'2
3 Omitted 疲 没③ &V蹤 6 Maximum (8charact- ers)
4 90rmore characters
凸暮'l
〔､Test Case Table〕
actorTest
Case
A C⊥D
Command
input
location
ACNAME operand
Entry
method
Parenthe-
S 1 S
aC C e S S n aれe
length
Test 宙 ① 宙咤 ①
Case1 彦W&ﾖ匁 ﾂ Spelled out &V7Fﾇ 7 V6貿坊B Minimum
Test 宙蔗 ② 宙 (冨)
Case2 彦W&ﾖ匁 ﾂ Mnemonic &V7Fﾇ 7 V6貿坊B Maximum
Test ① ②
Cases F6 Spelled 薈T7Fﾇ Ⅰnter-
out V6貿坊B mediate
I l l
㊨ : State number i n the test factor table

6. TEST FACTOR ANALYSI S
PROCESS
㊨ Factor Analysis
e to ExtTaet Input Factors
義 to Extract Environmental Factors
魯Analogy of Associated Factors
義 from One Key耶rd(Factor) to Another Key耶Td
療state Analysis
義 for Factors lndieating Numerie Values
I
e.蛋. `Equiv-alenee Partitioning ,
IBoudary Value Analysis'.
l for Factors Specifying a Selection Format
辛
e･g･ DSNAME(ifH盲_name恒*･ file-name,default,
none of above
l for FaetoTS in Generic Name Format
e.g. DASD亡)all types of DASD

7. TEST FACTOR ANALYSI S
METHODS
TestFactorAnalysis
Factor Analysis
Single Factor Analysis
義 to extract single factor
Analogical Factor Analysis
義 to extract associated faetoTS
State Analysis
Logical State Analysis
義 Equivalence partitioning
l Boundary value analysis
Factual State Analysis
DASD
operands of Commands
i
Z
H
u
割
O
e
⊥L
S
p
e
e
n
U
r
ハ
L
y
C
⊥
l
b
a
l
l

8. TEST KNO闇■LEDGE
〔Assoeiative Knowledge〕
Fi 1 e

9. UTI LIZATION OF
TEST KNOWLEDGE
Test factor
analysis screen
ど:ど 1
State-1
State-2
State-3 白
State-4
≡
State-1
State-2 白
State-3 ｢
State-4 亦 I 亦
〔Simple
Knowledge]
Te st factor
database
〔Associative
Knowledge]
α i
State-1
State-2 "
State-3 2
State-4
I
1
2
つ
ノ
4
b
b
b
b

10. STANDARD OF
TEST CASE GENERATI ON
1 0RTHOGONAL ARRAY 〔The Design of Ex,eriments〕
The same number of Combination of states
bet腎een any t耶faetors.
義 COMBINATION TABLE
At least one Combination of states
between any t耶faetors.

11. APPLICATION OF
THE COMBINATION TABLE
〔Test factor table〕
A C ﾂ
1 bl ﾂ
2 &#$3" 剴" a2
ヨ 2 剪
Standardized 劔剪
intotwostates
〔Test case table〕
A C
T1 bl
T2 " bl "
T3 ﾂ B23 "
T4 " B23 ﾂ
Combinationtable
applied(3factors)
ー123
T1
T2 "
T二ヨ )? ｨ b
T4 #
A C
Tl ﾂ bl ﾂ
T2 " bl "
T3 ﾂ b2 "
T4 ﾂ b3 "
T5 " b2 ﾂ
T6 " b3 ﾂ
Combinationtable
applied(1factor)
toeachB23
I1
T1 T2 R

12. TEST CASE DESIGN
SUPPORT SYSTEM
Test factor
F=
TESTCASEDESⅠ甜 剳B
SUPPORTSYSTEM
Testknowledge
accumulation
I 剩
function
tabl e
Test case
tabl e
ドニ
Test knowledge
search
functi on
Test factor
tableediting
andmanagement
function
Test case
generation
functi on
Combi nati on
tabl e
Testcase
tableediting
andmanagement
ftlnCtion
Test factor analysis
Testing
personnel
二二二手_
processing
Testing
personnel
筈LjOJ
Test case generation
α
Testcasel ﾂ b2
Testcase2 " b3
■､′ ′､′
Testing
personnel
三三_
:
:
:
:
J
J

13. CONCLUSI ONS
〔EXPECTED IMPROVEMENTS〕
◎Prevention of the ommission of factors and states
⑳Prevention of the ommission of test factor
eombination
⑳ E血eational effects for ne野membeTS
lFUTURE ENHANCEMENTS]
㊨ Automatic accumulation of Test Knowledge from
test faetoT table database
㊨ Selective usage of Test Kno酎edge by dynamic Change
of links between associated factors
⑳ Automatic interpretaion of external specifications
by formalizing them
External
specifications
汀]〒
⊂二つ
TestCaseDesi紳
SupportSystem
Formalizing
Te st Case
Tabl e
Te st Factor
Tabl e