Modelling

1
Modelling

• Learning objectives
• Able to recognize problems of learning to model
  entities and cardinalities

• Data modelling
• Making a model
• Presentation of student work
• Comparing object, process and data modelling
• Presentation
• Discussion

2
Example

• A database is going to include information about
  crimes, criminals and the commodities they steal.

• Make a data model of this part of the database!
• Groups of 2-3
• Plenary

3
ER modelling

• Boxing the world with words concept building as
  linguistic challenge for students of data
  modelling
• Christian Holmboe, 2002
• Method
• Three classes of high school students
• Age 18-19
• Second year of system development course
• Observations / interviews each week during

4
Types of concepts

• Spontaneous
• World ? Language ? Model
• Concrete
• Criminal, Commodity
• Semi-concrete
• Events, processes, types, aspects
• Crime, Crimetype, Judgement
• Scientific (relational)
• Model ? Language ? World
• First order relational phenomena
• Invented to resolve m-m relationships
• Crime Crimepart Crimetype
• Higher order relational phenomena
• Invented to resolve m-m relationships between
  relational phenomena
• Judgement Sentence Sentencetype

5
Relational concepts

• Students recursively generate more m-m
  relationships
• Students dont relate other entities to the
  relational ones

Commodity Name Value
Criminal Criminal-ID Name
Crime Crime-ID Timedate

• Participation
• Criminal-ID
• Crime-ID
• Crimetype

6
Getting the commodity in place

• R teacher, C,S students

7
Ogdens triangle
Adding database complexity
Subjective meaning (Sense, understanding,
intension)
is about
denotes
refers to
Expression
Table
Entity
Object (Extension)
Dog
refers to
refers to
refers to
8
Dogs have horns

• Students select a name for a relational entity
• The name is refers to objects which also have
  properties that are irrelevant for the phenomenon
  to be modelled
• The students add these properties to the entity
• Form_master is a teacher who is in charge of the
  same class of pupils for several years

9
Novice errors in conceptual DB designBatra and
Antony, 1994

• Method
• 2 lab experiments
• 31 / 29 students
• Halfway into first semester datamodelling
• Problem
• An agency making hotel room reservations
• Hotels, rooms, customers
• Analysis of verbal protocol (8 diff problems)

10
Results

• Students products
• 4 of 31 students had correct solution?
  difficulty
• Protocol analysis
• Procedural problems
• Literal translation of the task description
• Anchoring (stuck with initial solution)
• Incomplete knowledge (realize, but cant find
  error)
• Mistakes
• Incorrect arity
• Incorrect connectivity

11
Modelling

• Learning Object, Process and Data
  MethodologiesIris Vessey Sue Conger, 1994
• Method
• Process-tracing (Protocol analysis)
• Novice systems analysts
• 6 business students
• 3 different methodologies and 3 exercises
• Classification of Breakdowns
• Lot of statistics

12
Results
Process Data Object Easy Difficult

• Problem solvers prefer procedural strategies
• Declarative knowledge is assimilated first - then
  procedural knowledge.

Does the difficulty of object-oriented modelling
correspond with your experience?