Page 1
1© Amit Mitra & Amar Gupta
Topics
• Introduction to information Space• The concept of meta-object and meta-model• Introduction to Perspective, Classification and “The
Tyranny of Words”• Basic meta-object inventory
– Containers of normalized Knowledge– Also the basic components from which more
complex knowledge is configured
Page 2
04/18/23 13:47 page 2#
© Amit Mitra & Amar Gupta
BEHAVIOR
• RESPONSE TO A GIVEN STIMULUS– HIT METAL SHEET: it bends
– HIT GLASS SHEET: it breaks
• INVOLVES OBJECTS, EVENTS, CHANGE
• CHANGE INVOLVES TIME
• TECHNIQUES FOR REPRESENTING BEHAVIOR– BLACK BOX
» “INPUT-OUTPUT” VIEW
– NODE BRANCH
» “ERD TYPE” TECHNIQUES
Page 3
© Amit Mitra & Amar Gupta
DISCRETE CHANGE
• ATTRIBUTE VALUES & RELATIONSHIPS CHANGE IN RESPONSE TO DISCRETE EVENTS• CONSTRAINTS ON ENTITIES CHANGE IN RESPONSE TO DISCRETE EVENTS
Time slice(a single state of an
instance of an object)
OBJECT CLASS
Present
Past
V1
V2
V3
V4
V1
V2
V3
V4
Inst
ance
TimeTime
V1
V2
V3
V4
AN OBJECT CLASS IS ALSO AN INSTANCE OF AN OBJECT•What properties would the class normalize?
Page 4
© Amit Mitra & Amar Gupta
DISCRETE CHANGE
OBJECT CLASS
Present
Past
V1
V2
V3
V4
V1
V2
V3
V4
Inst
ance
Tim
e
V1
V2
V3
V4
Effect of hammer strike 1Effect of
hammer strike 2Effect of
hammer strike 3
Page 5
5© Amit Mitra & Amar Gupta
STATE OF AN OBJECT INSTANCE
1/8”
red
W
Instance
OBJECT CLASS(Glass Pane)
Status:Shattered (S) or Whole (W)
Color
Thickness
1/2”
blue
W
Instance
1/4”
green
S
Instance
Effect of Hammer Strike:Change status of “Whole” glass to
“Shattered”
Object Instances(Individual glass panes)
PROPERTIES OF OBJECT CLASS
Effect of hammer strike 3
Effect of hammer strike 2
Effect of hammer strike 1
Page 6
6© Amit Mitra & Amar Gupta
State Chart
Material Wholeness Color Thickness
Glass
Shattered
Whole
Hammer Strike (Whole)
PANE
Red
Blue
Page 7
7© Amit Mitra & Amar Gupta
State Chart
Material Wholeness Color Thickness
Glass Whole
Hammer Strike
PANE
Red
Blue
X
Cracked
Broken
Shattered
X
Page 8
8© Amit Mitra & Amar Gupta
State of a System
STATE OF INVENTORY SYSTEM
Lead time
STATE OF INVENTORY ITEM
Reorder Quantity
Quantity on Hand
Price
Quantity on Order
Under Scrutiny
STATE OF VENDOR
On Recommended Vendor List
Not recommended
FailedQA
Poor Performance
STATE OF VENDOR ITEM STATE OF VENDOR APPROVAL
PassedQA
Page 9
9© Amit Mitra & Amar Gupta
State Space
WEIGHT
TH
ICK
NE
SS
1/2 inch
1/8 inch
1 lb 2 lb
1/4 inch
11/2lb
Location of1/4 inch., 11/2 lbpane in this state space
Page 10
10© Amit Mitra & Amar Gupta
State Space with Ordinal axis
JANE’S PREFERENCE
CA
R S
IZE
Maximum Car Size
Minimum Car Size
MostLiked
LeastLiked
Second MostLiked
JANE’S PREFERENCE
CA
R T
YP
E(S
eque
nce
does
not
mat
ter.
Car
s ca
n be
ar
rang
ed in
any
ord
er a
long
this
axi
s)
Infiniti
Ford Explorer
MostLiked
LeastLiked
Second MostLiked
Honda Civic
A. Example of State Space when one axis maps to a quantitative domain and the other to a qualitative domain (Disjoint Lines)
B. Example of State Space when both axes map to qualitative domains (Disjoint Points)
STATE SPACE IS COLORED BLUE
(Sequence matters, but not the distance between the broken lines on this axis)
Page 11
11© Amit Mitra & Amar Gupta
THREE DIMENSIONAL STATE SPACE
Height
Age
Weight
Location in state space of an individual of a particular age, height and weight
•APPEARANCE OF STATE SPACE LOOK IF WE DID NOT CARE ABOUT THE AGE OF A PERSON?
•APPEARANCE OF STATE SPACE LOOK IF WE RESTRICTED THE AGE OF “PERSON” TO A SINGLE VALUE?
Page 12
12© Amit Mitra & Amar Gupta
TRAJECTORY IN STATE SPACE
Depth
Width
The cross section of the river (instance of an object) moves along this trajectory at a certain speed
Page 13
13© Amit Mitra & Amar Gupta
TRAJECTORY IN STATE SPACE
An object will move along a trajectory in State Space as its state changes with the passage of time
The object’s trajectory can be reinterpreted as a region in State Space (a static line in this case) when the time axis is added to its State Space
[When only discrete changes are considered, the region consists of a sequence of discrete points ( ) on the trajectory]
Depth
Width
The cross section of the river (instance of an object) moves along this trajectory at a certain speed
Location of a cross section of a river in state space at a particular time
Depth
Width
Time
Page 14
14© Amit Mitra & Amar Gupta
TRAJECTORY IN STATE SPACE
An object will move along a trajectory in State Space as its state changes with the passage of time
The object’s trajectory can be reinterpreted as a region in State Space (a line in this case) when the time axis is added to its State Space
[When only discrete changes are considered, the region consists of a sequence of discrete points ( ) on the trajectory]
Income
Borrowing
The firm (instance of an object)moves along this trajectory at acertain speed
Income
Borrowing
Time
Location of the firm instate space at aparticular time
Page 15
15© Amit Mitra & Amar Gupta
OBJECT
V1
V2
V3
E1
V1
V2
V3
E1
V1
V2
V3
E1
Instance Instance Instance
OBJECT CLASS
Time TimeTime
• OBJECT INSTANCE = SET OF VARIABLES– A single occurrence– Based on business meaning
• Person, place, category, concept or event relevant to the business • PROPERTY OF AN OBJECT (EXPANDED DEFINITION OF DATA ATTRIBUTES)
– A single meaning • Data Attribute/value• State/Relationship• Effect of event
• OBJECT CLASS = SET OF LIKE INSTANCES• TIME DIMENSION INTRINSIC TO BEHAVIOR OF THE OBJECT/CHANGES TO SPECIFIC PROPERTIES
Page 16
16© Amit Mitra & Amar Gupta
SUBTYPING
A
B ABsetintersection
AB is the set of objects that are members of both A and B.Multiple inheritance
A-B
B-A
CCA
setdifference
subsetof A
CA implies all members of C are also members of A, but not vice-versa.Inheritance (Data, behavior & constraints)
PERSON
MALE PERSON
Subtype of
AgeHeightWeight
(Inherited)AgeHeightWeight
GAS POISON
POISON GAS
Subtype of
Page 17
18© Amit Mitra & Amar Gupta
OBJECT GENERALIZATION & ROLE PLAYING
OBJECT
SUB-TYPE
.......PARTITION ....PARTITION
partitioned by partitioned by
partition of partition of
SUB-TYPE
SUB-TYPE
SUB-TYPE
(eg: organization)
e.g., Temporary organization(e.g.: Task force, Project team etc.)
e.g., Organization we do not own.
e.g., Organizations we own fully
e.g., Organizations in which we own the majority of shares
SUB-TYPE
e.g., Permanent. Organization(e.g.: Corporation, human resources department etc.)
PROPERTIES OF PARTITIONS
Irreducible fact: Subtypes are exhaustivelydefined in partition(Exhaustive Partition)
Irreducible fact: Subtypes are not exhaustivelydefined in this partition:Organizations in which we have minority shares are not shown(Non-exhaustive Partition)
Irreducible fact: Partitioning Criterion:Our ownership of organizations
Irreducible fact: Partitioning Criterion:Permanence of organizations
PARTITIONING CRITERIA
EXHAUSTIVITY
Non-exhaustive Partition Exhaustive Partition
*
e.g., Organization
Page 18
20© Amit Mitra & Amar Gupta
CONSTRAINTS BETWEEN SUBTYPES
Every instance in one subtype must be in the other subtype and vice-versa, I.e., the two sets are equal
Every instance in one subtype must be in the other subtype but not necessarily vice-versa
SUB-TYPE
A
Anothersubtype
OBJECT
PARTITION PARTITIONEvery instance of subtype A must also be an instance of subtype B
(not necessarily vice-versa) SUB-TYPE
B
Anothersubtype
Subtypes A and B are mutually exclusive even though they are in different partitions
SUB-TYPE
A
Anothersubtype
OBJECT
PARTITION PARTITIONAn instance of subtype A must not be an instance of subtype B
(mutually exclusive sets)SUB-TYPE
B
Anothersubtype
X
SUB-TYPE
A
Anothersubtype
OBJECT
PARTITION PARTITIONEvery instance of subtype A must also be an instance of subtype B
(not necessarily vice-versa)
SUB-TYPE
BEvery instance of subtype B must also be an instance of subtype A
(not necessarily vice-versa)Anothersubtype
Page 19
21© Amit Mitra & Amar Gupta
Subtyping Criteria
SUBTYPINGCRITERIA
Attributes EffectsRelationships Constraints
Constraintson Attribute
Values
Constraintson
Relationships
Guard Conditions
Constraintson Initial
Conditions
Initial Conditions
(Default State)
Constraintson History
Page 20
22© Amit Mitra & Amar Gupta
VARIATION INHERITANCE
PERSON
NON-PARENT
PARENT
PARENTHOOD PARTITION
Subtypes of Person
may be parent of 0 or more
parent of 1 or more(inherited from Person)
Features•Parenthood•Name•Age•Gender•Height•Weight
Features•Parenthood•Name•Age•Gender•Height•Weight
Features•Parenthood•Name•Age•Gender•Height•Weight
Inherited from
Person
Inherited from Person
EXCLUDEFEATUREFROM SUBTYPE
ADD FEATURETO SUBTYPEADD FEATURETO SUBTYPE
PERSON
NON-PARENT
PARENT
PARENTHOOD PARTITION
Subtypes of Person
parent of 1 or more(added to subtype) Inherited
from Person
Features•Name•Age•Gender•Height•Weight
Features•Name•Age•Gender•Height•Weight
Features•Name•Age•Gender•Height•Weight
Inherited from
Person
Parenthood(add feature to
subtype)
+
THE PRINCIPLE OF SUBTYPING BY ADDING INFORMATIONINCLUSION INHERITANCE
Page 21
23© Amit Mitra & Amar Gupta
Polymorphism is the quality of appearing in
several apparently different forms.
PARENT EMPLOYEE CUSTOMER ETC.
E.G. PERSON
HOP RUN ROLL ETC.
E.G. MOVE
MEETING TASK BIRTHDAY ETC..
E.G. EVENT
HEAT LIGHT KINETIC ETC..
E.G. ENERGY
BY ONES BY TWOS BY THREES ETC.
E.G. COUNT
GUIDELINE RULE DESCRIPTION ETC.
E.G. INFORMATION
●First identified by Christopher Strachey in 1967. ●Context specific behavior normalized by generalizing or subtyping objects.
– For example, the exact meaning of length depends on whether the object in question is a word or a room. “Word” and “Room” are parameters of length that fixes its meaning and properties more precisely than the generic concept of length:
•The length of a word is the number of letters in it, which can only be an integer•The length of a room may be any real number.
POLYMORPHISM
Page 22
24© Amit Mitra & Amar Gupta
OBJECTparameters
Adaptation through Inclusion Polymorphism
•Object=Frog; Move=Hop
•Object=Wheel; Move=Roll
•Object may be Frog or Wheel•Move may be Hop or Roll
Page 23
25© Amit Mitra & Amar Gupta
Kinds of Polymorphism
POLYMORPHISM
UniversalPolymorphism
AdhocPolymorphism
ParametricPolymorphism
InclusionPolymorphism
OverloadingPolymorphism
CoercivePolymorphism
Universal polymorphism is “true” polymorphism
Parametric polymorphism: Common behavior flows from domains and their mutual relationships. For example, the age of different kinds of objects such as people, documents or ideas may be computed by the following formula:
Age= Current Time - Time of creation
Current time and time of creation are two parameters needed to calculate “Age”, hence the term Parametric Polymorphism.
Inclusion polymorphism is the kind of polymorphism where subtypes inherit behavior. For example, Persons may speak, which means both male and female persons may speak.
Ad-hoc polymorphism is an artificial construct based on a somewhat unnatural and ad-hoc assignment of behavior or names to objects:
Overloading is the ability to use the same syntax for objects of different types, e.g. "+" for addition of complex numbers and integers or “length to compute the length of a room or a character string. Overloading reuses the name for a function) but requires different code for different types of objects. The function name is really a homonym (two different meanings with the same label) in overloaded polymorphism.
Coercive polymorphism occurs when an object instance is arbitrarily (and perhaps unnaturally) declared to belong to some subtype for a function. For example, special characters are assigned a sort sequence, just as numbers are, even though they do not have any natural sequence in which they must be arranged. Coercion also occurs when values in a nominal domain are arbitrarily assigned an order or magnitude; when differences or ratios between values in an ordinal domain are arbitrarily assigned a magnitude, or ratios of values in any but ratio scaled domains are compared.
OU
R FO
CU
S
OU
R FO
CU
S
Page 24
26© Amit Mitra & Amar Gupta
KINDS OF INHERITANCE
•Subtype Inheritance: Mutually exclusive subtypes inherit behavior of the parent class
•Extension Inheritance: State space of the subtype extends the state space of the parent into additional dimensions (has additional properties)
•Restriction Inheritance: Constraint is added to parent to restrict its state space in the subtype
– Lawful vs. Conceivable state space
•View Inheritance: Object is an instance of two or more different subtypes simultaneously and inherits properties and restrictions of all. Eg:
– Parent and Employee are two roles (subtypes/polymorphisms) of Person
– An individual may simultaneously have a pay rate and children if he or she is an employee and a parent at the same time
INHERITANCE
ModelInheritance
VariationInheritance
SubtypeInheritance
ViewInheritance
SoftwareInheritance
RestrictionInheritance
ExtensionInheritance
(See endnote on kinds of inheritance in your text book)
Page 25
© Amit Mitra & Amar Gupta
The Problem of Perspective
What you see or think depends on how you see or think
GOOD BAD
GOODBAD
DOES A UNIVERSAL
PERSPECTIVE EXIST
?
WITHOUT THE UNIVERSAL
PERSPECTIVE THERE WOULD BE NO SHARED
UNDERSTANDING
Page 26
30© Amit Mitra & Amar Gupta
Universal Perspective●Information and business services hub for semantic interoperability●Remember the Principle of Parsimony
–Generalized Classes and Interactions; Information Sparse●Remember the principle of subtyping by adding information
PERSPECTIVE
PERSPECTIVE
PERSPECTIVE
PERSPECTIVE
PERSPECTIVE
PERSPECTIVE
SHAREDPERSPECTIVE PERSPECTIVEPERSPECTIVE
Page 27
© Amit Mitra & Amar Gupta
GOOD BAD
GOODBAD
• TRUTH DEPENDS ON CONTEXT• CONTEXT ADDS INFORMATION
– PERSPECTIVE PROVIDES INFORMATION– IS A KIND OF OBJECT
CONTEXT 1
CONTEXT 2
Page 28
32© Amit Mitra & Amar Gupta
Synonyms and Homonyms
NAME(synonym
&homonym)
NAME(synonym)
NAME(synonym)
NAME(synonym)
NAME(synonym)
NAME(synonym)
NAME(synonym)
OBJECT 1(meaning)
OBJECT 2(meaning)
NAME(synonym)
NAME(synonym)
THE TYRANNY OF WORDS
NAME(homonym)
NAME(homonym)
OBJECT 1(meaning)
OBJECT 2(meaning)
Page 29
33© Amit Mitra & Amar Gupta
Primary Name vs Alias
ALIAS(synonym
&homonym)
ALIAS(synonym)
ALIAS(synonym)
ALIAS(synonym)
ALIAS(synonym)
ALIAS(synonym)
ALIAS(synonym)
OBJECT 1PRIMARY
NAME
OBJECT 2PRIMARY
NAME
ALIAS(synonym)
ALIAS(synonym)
Page 30
34© Amit Mitra & Amar Gupta
Aliases and Perspectives
PERSPECTIVE or CONTEXT
(Model)
PERSON/ or ORGANIZATION(Key stake holders:
Persons and Groups)
CONCEPT(object)
NAME(synonyms, i.e.,
Aliases)
Each concept must have a name, and may have many
Each name must be in a a context and maybe the same in many contexts
Each perspective must be heldby at least one person or organizationand may be held by many
Each name must be the name of at least one object, perhaps many
Each perspective must have at least one named concept, probably more
Each person or organization must hold at least one perspective, perhaps more
Page 31
35© Amit Mitra & Amar Gupta
DOMAIN
OBJECTINSTANCE
Described by exactly 1[Describes 1 or more]
Des
crib
ed b
y 1
or m
ore
[Pro
pert
y of
1 o
r m
ore]
EVENT
Triggers 1 or more[triggered by 1]
ATTRIBUTEVALUE
ATTRIBUTE
VALUE
Property of 1 or more[described by 1 or more]
Tak
es e
xact
ly 1
[sta
te o
f 1]
Described by exactly 1[Describes 1 or more]
STATEChange 1
[changed by 0 or more]Set of 1 or more
[component of 1 or more]
Subtype of
point in 1 or more[set of 0 or more]
OBJECT CLASS
Set (dimension) of 1 or more[described by 1 or more]
(implied)State Space
Belongs to 1[set of 1 or more]
AGGREGATEOBJECT
EFFECT
(Multiple occurrencesof an object are distinctmembers of a list)
LIST(Multiple occurrencesof an object count as a
single member)
CLASS or SET Subtype of
Partition of[partitioned by]
PERSPECTIVE
exists in 1 or more[pattern of 0 or more]
SequencedPattern
(Sequenced vs.Unsequenced Partition)
SequencedPattern
Relate 1 or more (relationship)
Subtype of 1 or more Represent 1 or more
Subtype of
OBJECTINSTANCE OBJECT
INSTANCE
Subtypeof
Subtype of
The Metamodel of Object
Page 32
© Amit Mitra & Amar Gupta
As a juggler plays many a part, wears many a guiseAnd doffs his mask when the show is done
So is our creator one; the only one- Guru Arjan Singh of Sikhism in the holybook, Adi Granth
M ET A OB JE C T
D YN A MI CRE LA T ION S HI P
(Proc es s )
ST A T I CRE LA T ION S HI P
RE L AT ION S HI PDOM A I N
Q UA L IT A TIV EDOM A I N
QU A NT ITA T IV EDOM A I N
D I FFE RE NC ES CA LE DDO MA I N
R A TI OS CA LE DDOM A I N
N OM INA LDO MA I N
O RD INA LDOM A I N
OB J E C TP R OP E RT Y
I r red uc ib le fac tin v ol ving tw o o r
mo re o bje ct s ,or t h e s am e
o bje c t a td iff e ren t t ime s
I rred uc ib le fac ta bou t t im i n g o rse que nce
T i m ein dep end en tI r red uc ib le fac t
A s ser ti ons ab ou tm eas ur eme nt an d
c la ss if ic a ti o n o f ob jec tb eh avio r c om mo n t om ul tip le pr ope r ti e s o f
ob j ec t s
I rre duc ib lefac t a bou tcla s si fic a tio nonl y
Ir r edu ci b le f ac t sa bo ut o rde r o rra n kin g o fob jec ts
Ir re du cib l e fac ts ab ou tma gni t ud es of d iff e ren c e sbe t we en obj ec t s
Ir r edu cib l e fac tsa bou tm ag nit ude s o fr a tio s b etw ee nobjec ts
UNI T O FM E AS UR E
( U OM )
F OR MA T
Ir r edu cib le fac tsa bo ut U ni t s o fm e asu rem en t
Ir r edu cib le f ac t sa bo ut f orm ats f o rp res en t ing ob jec t
pr ope rt ie s toob s er ver s
(hu ma n o rm ec h an ica l )
Is o la te dirre duc ib lefac t a bou ta s ing leob jec t
E V EN T
(Te mp ora lO c cu rre nce )
Ir re duc ib lefac t a bou tt im in g o f /t rigg er s f o rb eha vi o r
AT T RIB UTE
P O S S IB L EP A R T IC IP A T IO N
IN ARE LA T ION S HI P
EF F EC T
S UB T YP E
S T AT E
C olle c tio n o fI rre duc ib lef ac t s a bou t
th e c o nd itio no f an ob jec t
ins tan ce at amom e nt in
t im eS T AT ESP A C E
Th e s et o f a llp os s ible
co ndi t ion s o fan ob ject
R e pos ito ry o fun sha re dirre duc ib lefac ts
Re pos ito ry o fsha re dirre duc ib lefac ts
S UP E RT YP E
Ir r edu cib le f ac t ab ou t th eex ist e nce of a s ing le pro per ty o fan ob jec t ins tan ce at a m om en t
in t im e
Ir redu c ib le fact th epo s s ibilit y o f an ob jec t
in s ta nc e 's inv olv eme nt i n as pe cif ic r e la t ion s hip
OB J E C TCL AS S
S e t o f ob jec tin s tan ce sw ith s om esha re db eha vio r
OB J EC TIN S TA NC E
I rre duc ib lefac t a bou tt he ex is te nc eof a c on c ep to r it e m for afin i te i nt e rv a lof t im e
OB J EC TP A RT I T IO N
C ONS TR AIN T
C ri te r ia for d if fere ntia t in gs ub se t s o f ob jec t
ins tan c e s in an ob jec tc las s
I rre duc ib lef a ct a bo ut alim ita t io n
TR AJE CT OR YIN S T A TE
SP A C E
S E T O FP O S S IB L E
TR A JE CT ORI ES IN S T AT E
SP AC E
S e t of a l l p os s ib lec h an ges in st a te o f a nob jec t in s ta nc e
I r red uc ib le fac tab ou t th e p os s ib lec ha nge in th es tate o f an ob jectf rom on e m om en tto ano th er inre sp ons e t o a ne v en t
Th e s et o fac tua l
ch an ges ins tat e o f a n
ob jec tin s ta nc e
A GG RE G A TEOB JEC T
Ir r edu ci b le f ac t sa bo ut Coll ec t ion s
a nd st r uct ure s
NA ME
A l ter na t iv enam e s
UNI V E R S A LPE RS P E C TIV E
P E RS P EC T IV E
OT HE RP E RS P EC T IV E
M ode l:Se t o f
s tru c tu re di nt e rc o nn ect e
d c o nc epts
U niv ers a llys har ed c o nc eptsa nd me an ing s
S pec ia l iz e dc on cep ts an dme an i ng s
SY N ON YM H OM ON YM
D i f fe ren tN am es fo rthe sa m eob j ec t c las s
The sa m eNa me fo rd if fe ren tob jec tc l as se s
(Sta tes o f Name)
AL I AS C ON C EP TI D
Ob j ec t C l as sIde nt i fi e r
SU B TY P IN GRE L AT ION S HI P
I rre duc ib lef ac t s a bou texis ten c e of as ub t yp e
Page 33
37© Amit Mitra & Amar Gupta
What We Have Covered• The concepts of object class and object instance
– The fundamental containers of reusable knowledge – Also the fundamental meta-component from which other components of
knowledge are forged
• The concept of meta-object and meta-model
• State and state space
• Subtyping Partitions and Polymorphism
• Shared Understanding, Perspective, Classification and “The Tyranny of Words”
• A basic meta-object inventory– Containers of normalized Knowledge– Also the basic components from which more complex knowledge is configured