Chapter 7

 Requirements Modeling: Flow, Behavior,

Patterns, and WebApps
Slide Set to accompany
Software Engineering: A Practitioner’s Approach, 7/e
by Roger S. Pressman

Slides copyright © 1996, 2001, 2005, 2009 by Roger S. Pressman

For non-profit educational use only

May be reproduced ONLY for student use at the university level when used in conjunction
with Software Engineering: A Practitioner's Approach, 7/e. Any other reproduction or use is
prohibited without the express written permission of the author.

All copyright information MUST appear if these slides are posted on a website for student
use.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 1
Requirements Modeling Strategies
 One view of requirements modeling, called structured
analysis, considers data and the processes that transform
the data as separate entities.
 Data objects are modeled in a way that defines their
attributes and relationships.
 Processes that manipulate data objects are modeled in a
manner that shows how they transform data as data
objects flow through the system.
 A second approach to analysis modeled, called object-
oriented analysis, focuses on
 the definition of classes and
 the manner in which they collaborate with one another to
effect customer requirements.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 2
Flow-Oriented Modeling

 Represents how data objects are transformed at they

move through the system
 data flow diagram (DFD) is the diagrammatic form that
is used
 Considered by many to be an “old school” approach, but
continues to provide a view of the system that is unique
—it should be used to supplement other analysis model
elements

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 3
The Flow Model
Every computer-based system is an
information transform ....

computer
input based output
system

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 4
Flow Modeling Notation
external entity

process

data flow

data store
These slides are designed to accompany Software Engineering: A Practitioner’s Approach,
7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 5
External Entity
A producer or consumer of data

Examples: a person, a device, a sensor

Another example: computer-based

system

Data must always originate somewhere

and must always be sent to something

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 6
Process
A data transformer (changes input
to output)

Examples: compute taxes, determine area,

format report, display graph

Data must always be processed in some

way to achieve system function

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 7
Data Flow

Data flows through a system, beginning

as input and transformed into output.

base
compute
triangle area

height area

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 8
Data Stores
Data is often stored for later use.

sensor #
sensor #, type,
look-up location, age
sensor
report required data
type,
location, age
sensor number

sensor data

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 9
Data Flow Diagramming: Guidelines
 all icons must be labeled with meaningful
names
 the DFD evolves through a number of levels
of detail
 always begin with a context level diagram
(also called level 0)
 always show external entities at level 0
 always label data flow arrows
 do not represent procedural logic

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 10
Constructing a DFD—I
 review user scenarios and/or the data
model to isolate data objects and use a
grammatical parse to determine
“operations”
 determine external entities (producers
and consumers of data)
 create a level 0 DFD

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 11
Level 0 DFD Example
processing
user request requested
video
digital signal
video monitor
processor
video
source NTSC
video signal

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 12
Constructing a DFD—II
 write a narrative describing the transform
 parse to determine next level transforms
 “balance” the flow to maintain data flow
continuity
 develop a level 1 DFD
 use a 1:5 (approx.) expansion ratio

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 13
The Data Flow Hierarchy
a b
x P y level 0

a c p2
p1
f

p4 b
d 5
p3 e g
level 1

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 14
Flow Modeling Notes
 each bubble is refined until it does just
one thing
 the expansion ratio decreases as the
number of levels increase
 most systems require between 3 and 7
levels for an adequate flow model
 a single data flow item (arrow) may be
expanded as levels increase (data
dictionary provides information)

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 15
Process Specification (PSPEC)

bubble

PSPEC
narrative
pseudocode (PDL)
equations
tables
diagrams and/or charts

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 16
DFDs: A Look Ahead

analysis model
Maps into
design model

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 17
Control Flow Modeling
 Represents “events” and the processes that
manage events
 An “event” is a Boolean condition that can be
ascertained by:
• listing all sensors that are "read" by the software.
• listing all interrupt conditions.
• listing all "switches" that are actuated by an operator.
• listing all data conditions.
• recalling the noun/verb parse that was applied to the
processing narrative, review all "control items" as
possible CSPEC inputs/outputs.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 18
Control Specification (CSPEC)
The CSPEC can be:
state diagram
(sequential spec)

state transition table

combinatorial spec
decision tables

activation tables

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 19
Behavioral Modeling
 The behavioral model indicates how software will
respond to external events or stimuli. To create the
model, the analyst must perform the following steps:
• Evaluate all use-cases to fully understand the sequence of
interaction within the system.
• Identify events that drive the interaction sequence and
understand how these events relate to specific objects.
• Create a sequence for each use-case.
• Build a state diagram for the system.
• Review the behavioral model to verify accuracy and
consistency.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 20
 State Representations
 In the context of behavioral modeling, two different
characterizations of states must be considered:
 the state of each class as the system performs its function
and
 the state of the system as observed from the outside as the
system performs its function
 The state of a class takes on both passive and active
characteristics [CHA93].
 A passive state is simply the current status of all of an
object’s attributes.
 The active state of an object indicates the current status of
the object as it undergoes a continuing transformation or
processing.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 21
State Diagram for the ControlPanel Class
t ime r < lo cke d Time

t ime r > lo cke d Time locke d

p as s wo rd = in co rre ct
& n u mb e rOfTrie s < maxTrie s

re ad in g comparin g n u mb e rOfTrie s > maxTrie s

ke y h it
p as swo rd
e n t e re d do: va lida t e Pa s s w ord
p ass wo rd = co rre ct

s e le ct in g

act iv at io n s u cce s sfu l

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 22
The States of a System
 state—a set of observable circum-
stances that characterizes the behavior
of a system at a given time
 state transition—the movement from one
state to another
 event—an occurrence that causes the
system to exhibit some predictable form
of behavior
 action—process that occurs as a
consequence of making a transition

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 23
 Behavioral Modeling
 make a list of the different states of a system
(How does the system behave?)
 indicate how the system makes a transition
from one state to another (How does the
system change state?)
 indicate event
 indicate action
 draw a state diagram or a sequence diagram

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 24
 Sequence Diagram

h o meo wn er co n t ro l p an el syst em sen

s enso
sorsrs

syst em read in g
A
read y
p asswo rd en t ered

req u est lo o ku p
co mp arin g

resu lt

pa s s word = c orre c t
num be rOfTrie s > m a xTrie s req u es t act ivat io n
lo cked

t imer > lo cked Time

A

select in g

act ivat io n su ccessfu l act ivat io n su ccessfu l

Figure 8 .2 7 Se que nc e dia gra m (part ia l) for Saf e Hom e s e c urit y func t ion

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 25
Writing the Software Specification
Everyone knew exactly
what had to be done
until someone wrote it
down!

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 26
Patterns for Requirements Modeling
 Software patterns are a mechanism for capturing
domain knowledge in a way that allows it to be
reapplied when a new problem is encountered
 domain knowledge can be applied to a new problem
within the same application domain
 the domain knowledge captured by a pattern can be
applied by analogy to a completely different application
domain.
 The original author of an analysis pattern does not
“create” the pattern, but rather, discovers it as
requirements engineering work is being conducted.
 Once the pattern has been discovered, it is documented

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 27
Discovering Analysis Patterns
 The most basic element in the description of a
requirements model is the use case.
 A coherent set of use cases may serve as the
basis for discovering one or more analysis
patterns.
 A semantic analysis pattern (SAP) “is a pattern that
describes a small set of coherent use cases that
together describe a basic generic application.”
[Fer00]

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 28
An Example
 Consider the following preliminary use case for software
required to control and monitor a real-view camera and
proximity sensor for an automobile:
Use case: Monitor reverse motion
Description: When the vehicle is placed in reverse gear, the
control software enables a video feed from a rear-placed
video camera to the dashboard display. The control
software superimposes a variety of distance and orientation
lines on the dashboard display so that the vehicle operator
can maintain orientation as the vehicle moves in reverse.
The control software also monitors a proximity sensor to
determine whether an object is inside 10 feet of the rear of
the vehicle. It will automatically break the vehicle if the
proximity sensor indicates an object within 3 feet of the rear
of the vehicle.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 29
An Example
 This use case implies a variety of functionality that
would be refined and elaborated (into a coherent set of
use cases) during requirements gathering and modeling.
 Regardless of how much elaboration is accomplished,
the use case(s) suggest(s) a simple, yet widely applicable
SAP—the software-based monitoring and control of
sensors and actuators in a physical system.
 In this case, the “sensors” provide information about
proximity and video information. The “actuator” is the
breaking system of the vehicle (invoked if an object is
very close to the vehicle.
 But in a more general case, a widely applicable pattern is
discovered --> Actuator-Sensor

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 30
Actuator-Sensor Pattern—I
Pattern Name: Actuator-Sensor
Intent: Specify various kinds of sensors and actuators in an embedded system.
Motivation: Embedded systems usually have various kinds of sensors and actuators. These sensors
and actuators are all either directly or indirectly connected to a control unit. Although many of the
sensors and actuators look quite different, their behavior is similar enough to structure them into a
pattern. The pattern shows how to specify the sensors and actuators for a system, including attributes
and operations. The Actuator-Sensor pattern uses a pull mechanism (explicit request for information)
for PassiveSensors and a push mechanism (broadcast of information) for the ActiveSensors.
Constraints:
Each passive sensor must have some method to read sensor input and attributes that represent the
sensor value.
Each active sensor must have capabilities to broadcast update messages when its value changes.
Each active sensor should send a life tick, a status message issued within a specified time frame, to
detect malfunctions.
Each actuator must have some method to invoke the appropriate response determined by the
ComputingComponent.
Each sensor and actuator should have a function implemented to check its own operation state.
Each sensor and actuator should be able to test the validity of the values received or sent and set its
operation state if the values are outside of the specifications.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 31
Actuator-Sensor Pattern—II
Applicability: Useful in any system in which multiple sensors and actuators are present.
Structure: A UML class diagram for the Actuator-Sensor Pattern is shown in Figure 7.8.
Actuator, PassiveSensor and ActiveSensor are abstract classes and denoted in italics. There are
four different types of sensors and actuators in this pattern. The Boolean, integer, and real
classes represent the most common types of sensors and actuators. The complex classes are
sensors or actuators that use values that cannot be easily represented in terms of primitive data
types, such as a radar device. Nonetheless, these devices should still inherit the interface from
the abstract classes since they should have basic functionalities such as querying the operation

states.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 32
Actuator-Sensor Pattern—III
Behavior: Figure 7.9 presents a UML sequence diagram for an example of the Actuator-Sensor
Pattern as it might be applied for the SafeHome function that controls the positioning (e.g., pan,
zoom) of a security camera. Here, the ControlPanel queries a sensor (a passive position sensor)
and an actuator (pan control) to check the operation state for diagnostic purposes before reading or
setting a value. The messages Set Physical Value and Get Physical Value are not messages between
objects. Instead, they describe the interaction between the physical devices of the system and their
software counterparts. In the lower part of the diagram, below the horizontal line, the PositionSensor
reports that the operation state is zero. The ComputingComponent then sends the error code for a
position sensor failure to the FaultHandler that will decide how this error affects the system and
what actions are required. it gets the data from the sensors and computes the required response for
the actuators.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 33
Actuator-Sensor Pattern—III
 See SEPA, 7/e for additional information on:
 Participants
 Collaborations
 Consequences

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 34
Requirements Modeling for WebApps
Content Analysis. The full spectrum of content to be provided by
the WebApp is identified, including text, graphics and images,
video, and audio data. Data modeling can be used to identify
and describe each of the data objects.
Interaction Analysis. The manner in which the user interacts with
the WebApp is described in detail. Use-cases can be
developed to provide detailed descriptions of this interaction.
Functional Analysis. The usage scenarios (use-cases) created as
part of interaction analysis define the operations that will be
applied to WebApp content and imply other processing
functions. All operations and functions are described in detail.
Configuration Analysis. The environment and infrastructure in
which the WebApp resides are described in detail.

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 35
When Do We Perform Analysis?
 In some WebE situations, analysis and design
merge. However, an explicit analysis activity
occurs when …
 the WebApp to be built is large and/or complex
 the number of stakeholders is large
 the number of Web engineers and other contributors
is large
 the goals and objectives (determined during
formulation) for the WebApp will effect the business’
bottom line
 the success of the WebApp will have a strong
bearing on the success of the business

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 36
The Content Model
 Content objects are extracted from use-cases
 examine the scenario description for direct and
indirect references to content
 Attributes of each content object are identified
 The relationships among content objects
and/or the hierarchy of content maintained by a
WebApp
 Relationships—entity-relationship diagram or UML
 Hierarchy—data tree or UML

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 37
Data Tree
Ma rke t ingDe s c ript ion

Phot ograph
pa rt Numbe r

Te c hDe s c ript ion

pa rt Na me

c ompone nt Sc he ma t ic
pa rt Type

Vide o
de s c ript ion

pric e
Whole s a le Pric e

Ret a ilPric e

Fig u re 1 8 . 3 Da t a t re e fo r aS a fe Ho m e c o m p o n e n t

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 38
The Interaction Model
 Composed of four elements:
use-cases
 sequence diagrams

 state diagrams

 a user interface prototype

 Each of these is an important UML notation

and is described in Appendix I

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 39
Sequence Diagram
:Produc t :Billof FloorPla n BoM
:Room :FloorPla n
Compone nt Ma t e ria ls Re pos it ory Re pos it ory

new cust o mer

d e sc rib e s
ro o m *
p la c e s ro o m
in f lo o r p la n

sa v e f lo o r p la n c o n f ig u ra t io n

se le c t s p ro d u c t c o m p o n e n t *

a d d t o Bo M

sa v e b ill o f m a t e ria ls

Figure 18.5 Se que nce diagram for use -case :se le ct Saf e Hom e com pone nt s

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 40
State Diagram
Va lida t ing use r Se le ct ing use r a ct ion
use rid se le c t ot he r func t ions
syst e m st a t us=“input re a dy” va lida t e d syst e m st a t us=“link re a dy”
se le ct “log-in” displa y msg = “e nt e ruse rid” displa y: na viga t ion choice s”
displa y msg =“e nt e r pswd”
pa ssword va lida t e d
e nt ry/ va lida t e d use r
e nt ry/ log-in re que st e d
n e w cu st o m e r do: link a s re quire d
do: run use r va lida t ion
e xit / use r a ct ion se le ct e d
e xit / se t use r a cce ss swit ch

c ust omiz a t ion c omple t e

se le c t e -c omme rc e (purc ha se ) func t iona lit y

se le c t c ust omiz a t ion func t iona lit y

ne xt se le ct ion
Sa ving floor pla n
Cust omiz ing se le ct de script ive
cont e nt sy st e m st a t us=“input re a d y”
syst e m st a t us=“input re a dy” De fining room se le ct de script ive displa y: st ora ge indica t or
displa y: ba sic inst ruct ions cont e nt
room be ing de fine d syst e m st a t us=“input re a dy” e nt ry/ flo or pla n sa ve se le cte d
displa y: roomde f. window do: st ore flo or pla n
e nt ry/ va lida t e d use r e xit / sa v e comple t e d
do: proce ss use r se le ct ion
e nt ry/ roomde f. se le ct e d
e xit / cust omiz a t ion t e rmina t e d a ll rooms do: run room que rie s
de fine d do: st ore room va ria ble s
e xit / room comple t e d

Building floor p la n se le ct de script iv e

se le ct sa ve floor pla n
cont e nt
syst e m st a t us=“input re a dy”
se le ct e nt e r room in floor pla n displa y: floor p la n window

e nt ry/ floor pla n se le cte d

do: inse rt room in pla ce
do: st ore floor pla n va ria ble s
e xit / room inse rtion comple t e d
room inse rt ion comple t e d

Figure 1 8 .6 Pa rt ia l s t a t e dia gra m for n e w c u s t o m eintr e ra c t ion

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 41
The Functional Model
 The functional model addresses two
processing elements of the WebApp
 user observable functionality that is delivered by the
WebApp to end-users
 the operations contained within analysis classes that
implement behaviors associated with the class.
 An activity diagram can be used to represent
processing flow

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 42
Activity Diagram
init ialize t ot alCos t

no c omp one nt s re ma in o nBoMList c omp one nt s re ma in o n Bo MList

invoke g et price and

calcShipping Cost quant it y
ret urns:
shipping Cost

lineCos t =
price x quant it y
invoke
det erm ineDiscount

ret urns : discount add lineCos t t o

t ot alCos t

d isc o unt >0

t ot alCost=
t ot alCost - discount

d isc oun t < = 0

t axTot al=
t ot alCost x t axrat e

priceTot al =
t ot alCost + t axTot al
+ shipping Cos t

Figure 1 8 .7 Ac t ivit y dia gra m for c o m p u t e Pr i (c )e o p e r a t i o n

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 43
The Configuration Model
 Server-side
 Server hardware and operating system environment
must be specified
 Interoperability considerations on the server-side
must be considered
 Appropriate interfaces, communication protocols and
related collaborative information must be specified
 Client-side
 Browser configuration issues must be identified
 Testing requirements should be defined

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 44
Navigation Modeling-I
 Should certain elements be easier to reach (require
fewer navigation steps) than others? What is the priority
for presentation?
 Should certain elements be emphasized to force users
to navigate in their direction?
 How should navigation errors be handled?
 Should navigation to related groups of elements be
given priority over navigation to a specific element.
 Should navigation be accomplished via links, via search-
based access, or by some other means?
 Should certain elements be presented to users based on
the context of previous navigation actions?
 Should a navigation log be maintained for users?

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 45
Navigation Modeling-II
 Should a full navigation map or menu (as opposed to a single
“back” link or directed pointer) be available at every point in a
user’s interaction?
 Should navigation design be driven by the most commonly
expected user behaviors or by the perceived importance of the
defined WebApp elements?
 Can a user “store” his previous navigation through the WebApp
to expedite future usage?
 For which user category should optimal navigation be
designed?
 How should links external to the WebApp be handled?
overlaying the existing browser window? as a new browser
window? as a separate frame?

These slides are designed to accompany Software Engineering: A Practitioner’s Approach,

7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman. 46