Вы находитесь на странице: 1из 12

Attributes of Information System Development

Author(s): Hugh F. Juergens


Source: MIS Quarterly, Vol. 1, No. 2 (Jun., 1977), pp. 31-41
Published by: Management Information Systems Research Center, University of Minnesota
Stable URL: http://www.jstor.org/stable/249166
Accessed: 01-10-2015 03:32 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

Management Information Systems Research Center, University of Minnesota is collaborating with JSTOR to digitize,
preserve and extend access to MIS Quarterly.

http://www.jstor.org

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

While numerous computer-based information


Attributes of systems have been developed and are in operation
there is no unified or commonly accepted theory to
Information System guide the development process. If there were such a
theory, I suspect that a higher proportion of these
Development systems would be deemed successful [see 16, 20]
and the systems would be more readily adaptable
By: HUGH F. JUERGENS to change with a corresponding reduction in the
high cost of maintenance.
This paper draws upon a variety of sources in an
effort to integrate much of the current thinking on
the information system development process. A
general model of the development process is
presented to provide increased understanding of
the effect that this process has upon the
characteristics of the resulting system. The concept
of decoupling is introduced as a framework for
guiding the development process toward
achievement of the system's desired capabilities,
while recognizing and planning for the
accommodation of change.

Attributes of an
Information System
The criteria by which the attributes of a system can
be evaluated are presented first in order to provide
a perspective from which the system development
process can be evaluated. These criteria are user-
oriented and fall into the broad categories of
effectiveness and cost. The effectiveness attribute
has to do with the ability of the system to satisfy the
users' needs and gives an indication of the benefits
or value provided by the system. Cost, on the other
hand, constrains the development of the systems.
Primary attention in this article will be directed
toward the effectiveness with which the system
satisfies users' needs. Ideally, desired attributes of
the system are specified prior to the development
Abstract of the system and become objectives that are used
to guide the development process.
Computer-based information systems too often do not
satisfactorily meet the information needs of users and are
inflexible in accommodating change. Because change is
inevitable for any system, a framework for viewing the System effectiveness
information system development process which takes System effectiveress has three dimensions:
this into account is necessary. The concept of decoupling capability, stability, and adaptability [3].
as a framework which seeks to minimize the adverse
Capability refers to the ability of the existing
consequences of change is discussed in this article.
system to satisfy the information needs of the
Keywords: Development objectives, development methodology, operations and decision-making functions in the
development process information system structure, organization. As such, we can look at the outputs
systems approach, decoupling.
of the system and compare them with what is
Categories: 1.3, 2.4, 2.43, 3.5, 4.0 needed. Several aspects of capability include the

MIS Quarterly/ June 1977 31

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

ability of the system to perform its intended of the art. Even then, placing a value on improved
functions by providing the necessary output and decision-making or increased customer goodwill is
the relevance of the output in satisfying the needs difficult, if not impossible. Benefits associated with
of the users. The fact that "well functioning" the provisions for achieving stability and
systems have unsatisfied users, as described by adaptability are not realized until the changes
McKinsey and Company [16] and Powers and become imminent, and the danger still exists that
Dickson [20], gives rise to this distinction between the provisions are not adequate due to unforeseen
the two similar views of capability. For example, in circumstances. Yet, resources are expended to
the absence of user participation when defining the "ensure" the effectiveness of the system before its
content and format of reports, the development value is proven.
personnel may define the output in terms that may
be difficult for the users to understand, even It is a risky venture to provide for the dimensions
of capability, stability, and adaptability. This risk
though all the necessary information may be
has two distinct, opposite sides: excessive costs are
present. Other aspects of capability include the
incurred both by providing capabilities that are not
accuracy and timeliness of the system's outputs.
needed or by failing to provide required
Stability and adaptability refer tofuture capability capabilities. It is in the context of this uncertainty
and are intended to protect the organization's that the planning and development of an
information system investment. Stability is the information system takes place.
degree to which the capability of a system persists Conflict exists between the objectives themselves
through changes in the system's non-user and the means for achieving them. For example,
environment. In other words, if the capabilities of
there are several ways in which the cost of
the system remain intact through variations in the
workload, changes in the hardware and hard- developing a system can be reduced. One
alternative is to decrease one or more of the
software (e.g., operating system, compilers, data
dimensions of effectiveness, e.g., eliminating some
management systems, etc.), and changes or of the desired capabilities or reducing the
additions to other application programs and files,
documentation effort. Such means of achieving a
then the system may be said to be highly stable.
cost reduction may be unacceptable if the
Adaptability, on the other hand, is the degree to
which adjustments to the system can be made as a objectives for the system have already been set.
result of changes in the user environment. An They have often been used as a short-run measure
to cut costs in a project and are symptomatic of
example is the ease with which changes in the inaccurate initial estimates and an inflexible
information needs of the users can be incorporated
into the system. project control system [20].

Development Process
System Cost
Traditionally, the system development process
The attribute of cost refers to the expenditure of within the system life cycle has been described as a
resources during the development, operation, and series of chronological phases or stages [5, 18].
maintenance of the system in order to provide the Different authors have diverse labels for the stages
degree of desired effectiveness. The most visible and have varying conceptions of what must be
resource, capital, is manifested in the information done in each stage. Each of these approaches has
system organization through hardware, software, its own merits. A "new"series of stages, as such, is
personnel, and supplies. not presented in this paper. Viewing the
Theoretically, both the benefits and costs can be development process at the system level as
stated in terms of a single quantity, such as dollars, definition, design, and construction can draw these
whose value is to be optimized. It is the approaches together.
responsibility of the development personnel to
produce the optimum system. Unfortunately many
of the benefits and costs cannot be quantified or Definition
measured a priori. It is not possible to determine During the definition stage, the information
meaningfully if a system will adequately meet user system is placed into the organizational context. It
needs until after it is in use, given the present state is in this stage that the objectives or desired

32 MIS Quarterly/ June 1977

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

attributes of the system are defined and the control over the costs of production and increased
boundaries for the system are delineated. The responsiveness to customer demand. As shown in
result of the analysis forms a set of requirements Figure 2, it may be initially comprised of
which include a definition of the overall functions forecasting, production scheduling, purchasing,
to be performed, system inputs, system outputs, raw material inventory control, and finished goods
performance, and integrity. inventory control subsystems. Each of these
subsystems may be factored further into
subsystems based upon differences in function or
Design in the timing of inputs or outputs. Within each of
The design stage is concerned with translating the these subsystems or processing cycles, programs
requirements for the system into specifications for may be identified for purposes such as input
providing the needed capabilities. One common editing, file update, and reporting. Each of the
practice associated with system design, especially programs is subdivided further according to
when dealing with large-scale systems, is the functions performed such as input, output, and
factoring of the system into subsystems; each of type of calculation. Modules are further factored
these subsystems is in turn further factored into and refined into submodules until subroutines are
identified and finally the individual logic elements
subsystems. In this manner, the amount of detail
that must be analyzed, assimilated, and that eventually result in programming language
statements are identified.
synthesized at any one point in time is reduced. The
underlying logic of this approach is that each of the The arrows in Figures I and 2 identify the
subsystems is easier to handle than the system of interfaces between subsystems and represent the
which it is a part, the subsystems can be handled flow of data. At the upper levels, this data will, for
somewhat independently, and ultimately the the most part, be contained in files although each
subsystems can be recombined or joined in a arrow does not necessarily represent a single file
straightforward manner to form the complete [13, 14]. At the lower levels, data will be passed
system [10, 12, 13, 15]. from one routine to another in central memory.
The design process for a system consists of
understanding the context of the system with
respect to its implications for the super-system of
which it is a part: Construction
* factoring the system into more manageable The construction stage is concerned with
and somewhat independent subsystems, translating the system's specifications into the
* specifying the requirements to be means for providing the identified capabilities. At
accomplished by each subsystem, and each level in the hierarchical structure,
* specifying the interfaces between the construction consists of providing the interface
subsystems. between the subsystems and determining when
each subsystem should be executed.
Together with the interface specifications, the
design stake results in the definition stage for each There are two schools of thought regarding how
of the subsystems. Each of these subsystems construction should be integrated with the design:
becomes the system of concern when it goes top-down design with bottom-up construction, or
through its design stage. The factoring process top-down design with top-down construction.
continues with increased refinement at each level These two approaches are illustrated in Figures 3
and is increasingly aimed at the "mechanics" of and 4. In bottom-up construction, the lowest level
how the system will function in the hardware and subsystems are constructed first and tested. The
in manual procedures. It continues until the next higher level is concerned with providing the
system's objectives are refined into a form that can necessary logic to coordinate their execution. This
be understood by the computer, whether in a portion of the complete system constructed thus
higher level language, such as COBOL, or an far is then tested. Construction in this manner
assembly language. Figure 1 illustrates the proceeds up through the hierarchy until the entire
decomposition process. system has been completed [2, 8, 19].
A production control system, for example, might In top-down construction, the sequence of events is
begin with an overall objective of achieving better essentially reversed. Thus, at a given level in the

MIS Quarterly/ June 1977 33

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

Level 0:

Level 1:

Level 2: -

Level 3:

Figure 1. Hierarchical Network Structure

34 MIS Quarterly / June 1977

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

Raw Material Receipts


Shipments-Production ... Purchase Orders
Customer DemandControl _ Current Production Status
Customerg
Demand System Production Schedule
Engineering Specs R

al \
Maintain Maintain Prepare
Finished Raw urchase
Goods Material Purchase
Inventory Inventory Orders Orders

]r ^^^. \ \ T ICurrent
Forecast Schedule Status
Demand ,, Production
Demand __-- Production
lg /7~ ' Schedule
lg - F,'

Finished S
Raw Raw- '
Finished Material Material
Goods .. j On Hand Usa

Determine Monitor Current


Raw Current Production
hled MaterialOptOperations Status
ds IAvailabilit
and t

Determine Determine Determine


id - Finished - Total Mfg. Master - Production
Goods
st Requirements Requirements Schedule Schedule

Determine
_- / Parts
ng 1 E Capacity
Exposion
Availability
/
/ Raw \
/ Material Finished
/ Usage Goods

Calculate Maintain Current


Material e Work In - - Production
lUsa e Prncess Status

Work In Time
Process Usage

Figure 2. A Partial Example of the Hierarchical Network Structure


for a Production Control System

MIS Quarterly / June 1977 35

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

Level 0: Construct - O perate

Level 1:
I
Define 1 Design -_ Construct

Level 2: Define -^ Design


I

Level n: Define Design Construct


Construct

Figure 3. Top-Down Design With Bottom-Up Construction

Level 0: Define mDesign ,m Construct

Level 1:
I
Define ,Construct
Design

Level 2: Define - Design Construct

Level n:
\
Define
\
Design mm Construct m,Operate

Figure 4. Top-Down Design With Top-Down Construction

36 MIS Quarterly / June 1977

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

hierarchy the subsystems are defined and the will be done next. Checkpoints can be predefined:
interfaces and coordination are constructed and the development process can be viewed as a
tested before the subsystems themselves are chronological series of phases or stages [5]. Other
designed and constructed [2, 4, 19]. checkpoints may arise spontaneously: top-down
program design or stepwise refinement can be used
in program development [2, 4, 24, 25].
Attributes of the
Development Structure System/subsystem decoupling does not imply that
The structure of the system development process a system is independent of its subsystems. Quite the
can be viewed both as a hierarchy and a network, contrary, it is because of the dependence of a
as shown in Figure 1. There is a hierarchical system's effectiveness upon the effectiveness of its
relationship between a system and its subsystems subsystems that this separation should occur. This
and there is a network relationship among the allows development personnel to formulate
subsystems. explicitly the objectives for each subsystem. Such a
formalization forms a basis for discussion with the
users to confirm or revise the system objectives and
Vertical decoupling provides a mechanism for recording the many
details of the design. Additionally, control over the
The hierarchical relationship between a system and development process occurs through a comparison
its component subsystems may at times be of the objectives for each subsystem with its design.
clouded. For example, development personnel
may be satisfied with an intuitive feeling regarding The definition of the lower-level subsystems,
the objectives for a subsystem before proceeding
during the design of a higher-level subsystem,
down to another level in the hierarchy. If one resembles the "black box" concept in which the
proceeds in this manner, the link to the overall inputs and outputs of a subsystem are defined, but
system objectives becomes nebulous and the details for accomplishing the transformation
formalization of objectives and subsystem are left undefined. We should concentrate on
structure at the lower level becomes more difficult
specifying realistic objectives for subsystems at
because of the greater amount of detail that must
succeeding levels and not become embroiled in the
be assimilated. details of how the objectives will be accomplished.
Decisions regarding the actual functioning of a
If an effort on the other hand is made at each level system should be made only at the level in which
in the hierarchy to formalize the objectives and the the detail can be placed into perspective and can be
interface between the lower level subsystems, a understood and properly evaluated. In the
clearer distinction will exist between successive previously mentioned production control
levels. The extent to which this distinction exists is example, if a variation of the classic economic
referred to as the degree of system/subsystem or order quantity formula [6] is used, we do not need
vertical decoupling. A high degree of vertical to derive the specific formula until we design the
decoupling implies that there is a logical separation module that will calculate it. Prior to this point, all
between subsystems at adjacent levels in the that need be known is that the quantity will be
hierarchy from the standpoint that firm objectives determined by minimizing the appropriate costs.
are set forth for each subsystem prior to its
development. Hence, the definition of objectives It should be noted that no general guideline can be
for a higher level system exerts a binding influence established for the number of levels in the system
on the lower level subsystems. The capability of the structure. The factoring process should result in a
system will be enhanced if the developmental tractable set of subsystems for the development
efforts at each level are focused on the
personnel. The tractability depends upon the skill
accomplishment of a unified purpose. and experience of development personnel and the
scope and complexity of the system. Some insight
Tangible evidence of vertical decoupling in the into the problem of tractability for both the
development process takes the form of management of the development effort and for the
documentation and checkpoints. Periodically we actual development can be found in the literature
stop to summarize what has been done and what [2, 17, 23].

MIS Quarterly/ June 1977 37

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

Horizontal decoupling development process and in providing greater


flexibility in scheduling the development efforts.
A network structure, as shown in Figure 1, can be
Simplication of the development is achieved by
used to depict relationships among subsystems at a
reducing the amount of detail that must be
given level in the hierarchy. Horizontal or comprehended at any one point in time.
subsystem/subsystem decoupling refers to the Scheduling flexibility is achieved by permitting
degree of independence among subsystems in the both concurrent development and implementation
network. The rationale for this form of decoupling of subsystems so that partial capabilities within a
is primarily to enhance the stability and long-term plan can be achieved.
adaptability of the resulting system by isolating the
Two problems associated with both vertical and
impact of potential sources of change upon the
system. horizontal decoupling are the cost of the
decoupling mechanism itself (e.g., documentation
This independence can be achieved in two ways. and database management systems) and the
First, the number of interconnections among potential for suboptimization. The alternative to
subsystems can be minimized. Hence, need for decoupling, however, is to stretch people's ability
communication and coordination is reduced. This to perceive the masses of detail beyond their
can be done by an appropriate choice of subsystem
relatively narrow limits. We have to be content to
boundaries so that the number of interconnections tackle a large-scale problem a little at a time, until
is reduced, or by defining a special interface among
computers are used more effectively to automate
subsystems or clusters of subsystems. An example the development process [21, 22].
of this latter approach is the use of database
management systems. Data independence is the
term used in the database management literature Factors to Consider
for decoupling [1, 9, 11]. in the Implementation
A second means of achieving horizontal of Decoupling
decoupling is to minimize the degree of
interconnection. The subsystems should not make Viewing the development process in terms of the
system life cycle may lead the reader to the
assumptions about the internal workings of other conclusion that system development is sequential
subsystems to achieve a high degree of
in nature. That is, first the system proceeds
independence. Even though certain assumptions
through the definition stage, then the design stage.
may be valid at one time, changes to a subsystem the construction
and finally stage.
can impact other related subsystems. This means
of horizontal decoupling is aided partially by Chronologically, this may be true. However,
vertical decoupling and by specifying the functions recursion, parallelism, and iteration are inherent in
the development process.
to be performed by each subsystem and deferring
the decisions regarding the means for
accomplishing the intended functions. For
Recursion
example, Myers [19] has identified six degrees of Recursion in the development process is
interconnection between program modules exemplified by the fact that at each level in the
ranging from content coupling (lowest or worst development hierarchy, the overall methodology is
decoupling) whereby modules refer directly to the the same, although the orientation and details of
contents of other modules (e.g., to modify a implementing the methodology differ. For
program statement) to data coupling (highest example, the upper levels in the hierarchy are
decoupling) whereby all data elements input and directed more toward the overall problem to be
output of a module are passed as arguments and solved and hence are oriented to the users and
the arguments are not used to control execution involve interpersonal skills. The lower levels are
(e.g., flags). directed toward the computer and the physical
realization of the system in hardware and software.
Development structure overview The development of a subsystem at any given level
Besides enhancing the various dimensions of in the hierarchy is in general concerned with
system effectiveness, horizontal and vertical obtaining, understanding, translating, and
decoupling are also of benefit in simplifying the communicating information. Information is

38 MIS Quarterly/ June 1977

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

obtained about the plans of higher levels in the change during the development of the system. As
hierarchy, because it is these plans that set the we proceed further into development, insight on
objectives for the subsystems. The technology the part of the development personnel is gained
available to the information system sets into the properties of the system. Alternatives may
constraints with respect to what can be done. The be evaluated: inconsistencies, errors, omissions,
environment, which may include other existing and other desirable but unprovided capabilities.
information systems, may also set constraints on The users similarly through participation can be
the development. The people who will eventually expected to gain increased awareness of (1) their
be involved with the system, whether they are the information and decision making environments,
users of the output or preparers of the input, add and (2) the capabilities and limitations of
information that will help in further refining the computerized information processing [20]. These
system. Achieving an understanding through factors all indicate that it may be necessary to go
analysis goes hand-in-hand with obtaining the back a level in the hierarchy to refine the plans for
information; analysis of the information may point the lower level.
to the need for more information. Design consists
A high degree of vertical decoupling facilitates
of translating the results of the analysis into more
iteration between levels. Even though the
detailed and specific plans which form the
objectives are subject to change, a clear statement
objectives for the lower level subsystems. These of the objectives for each subsystem is necessary to
plans are then communicated to prior phases for track their evolution and signal any change.
review (feedback) and to subsequent phases for
Otherwise there is a risk of having implicit
further development (feedforward).
objectives for each subsystem that reflect
This indicates that the system designer, in addition assumptions on the part of the development
to addressing a current subsystem level, should personnel rather than having explicit objectives
think both up one level and down one level in the formulated in the context of the overall system. On
hierarchy. Looking at the higher level allows the the other hand, having a well documented but
system to be viewed in its context within a broader inflexible set of objectives can be just as
perspective. The designer should think down one dysfunctional as vague objectives. An
level because that is the object of his development unsatisfactory system may result by not taking
efforts. Often, lower levels will set constraints upon advantage of the insight gained through the
the activities at a higher level, such as the development efforts by the users and development
capabilities of existing hardware. personnel.

Some iteration is seen to be of a rather informal


Parallelism
nature: i.e., iteration that is carried out as the need
Parallelism in the development process indicates arises. On the other hand, some may be of a formal
that the subsystems, at any given point in time, nature: i.e., iteration that is carried out as the need
may be in different stages of completion; some arises. On the other hand, some may be of a formal
subsystems may be in the definition stage, others in to either continue, revise, or discontinue the
the design stage, and yet others in construction. project. While it may seem desirable to have a
This parallelism adds scheduling flexibility to the number of these checkpoints to maintain close
development process so that partial capabilities control, whatever parallelism exists in the
within a long term plan can be provided. development up to a checkpoint must halt and
Alternatively, devoting resources to different activities that lag behind must be allowed to catch
subsystems concurrently may reduce development up to the early finisher.
time. The degree of parallelism that can exist
depends upon the degree of vertical and horizontal The preceding discussion focused on iteration
decoupling that is present. between levels. Iteration is also experienced in the
development efforts of each subsystem. For
example, one must obtain information about the
Iteration subsystem before it can be analyzed. However, this
Iteration acknowledges the fact that the objectives analysis may point to the need for additional
of a system, and in turn the subsystem objectives, information. Similarly, analysis of the subsystem
may not be completely defined and may even precedes its design, but along with design comes

MIS Quarterly/ June 1, 1977 39

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

increased understanding of the subsystem which that could yield unexpected and undesirable
may point to the need for further analysis; this, in properties for the system. Further, decoupling
turn, may require additional information, etc. recognizes the inevitability of change and in
essence strives to minimize its adverse
The amount of iteration experienced in system
consequences by planning for change.
development depends upon the degree to which the
system fits into existing molds and the extent to
which development personnel recognize and are References
able to take advantage of the correspondence. For [1] ANSI/X3/SPARC Study Group on Data Base
example, developing a payroll system may involve Management Systems. "Interim Report," EDT, Vol. 7, No.
little iteration whereas a system to furnish a 2, (75-02-08), pp. 1-140.
[2] Aron, Joel D. The Program Development Process: The
manager with information needed to support Individual Programmer, Reading, Mass., Addison-
decision-making activities may require extensive Wesley, 1974.
iteration. [3] Ashenhurst, R. L. (ed.). "Curriculum Recommendations
for Graduate Professional Programs in Information
Systems," A report of the ACM Curriculum Committee on
Conclusion Computer Education for Management, Communications
systems are of the ACM, Vol. 15, No. 5 (May 1972), pp. 363-398.
Many current information
[4] Baker, F. Terry. "Chief Programmer Teams Management
unsuccessful in that they do not adequately satisfy
of Production Programming," IBM Systems Journal, Vol.
user needs and are inflexible when change is 11, No. 1 (1972), pp. 56-73.
required. Part of this problem occurs because [5] Benjamin, Robert I. Control of the Information System
information s, stem development is frequently Development Cycle, New York, Wiley, 1971.
viewed as strictly a technical activity and, as [6] Buffa, Elwood S. Operations Management: Problems
and Models, New York, Wiley, 1972.
designers, we often are in too much of a hurry to [7] Canning, Richard G. (ed.). "That Maintenance'Iceberg',"
delve into the details of the system and produce EDP Analyzer, Vol. 10, No. 10 (October 1972).
tangible evidence of our labor (e.g., computer [8] Canning, Richard G. (ed.)., "The Search for Software
Reliability," EDP Analyzer, Vol. 12, No. 5 (May 1974).
programs). CODASYL Systems Committee. Feature Analysis of
[9]
Computer science has recognized this problem for Generalized Data Base Management Systems, New York,
some time now. As a result, techniques such as Association for Computing Machinery, May 1971.
structured programming, modular programming, [10] Davis, Gordon B. Management Information
Systems: Conceptual Foundations, Structure, and
top-down program design, and HIPO increasingly Development, New York, McGraw-Hill, 1974.
are being accepted and used. All these techniques [11] Everest, Gordon C. "Managing Corporate Data
for forcing disciplined creativity incorporate the Resources: Objectives and a Conceptual Model of
Database Management Systems," Unpublished Ph.D
concept of decoupling and are oriented more
dissertation, University of Pennsylvania, 1974.
toward the lower levels in the development
[12] Johnson, Richard A.; Kast, Fremont E.; and Rosenzweig,
hierarchy. Reported benefits of these approaches James E. The Theory and Management of Systems, 3rd
include reduced development time and cost, edition, New York, McGraw-Hill, 1973.
increased quality of programs, and reduced [13] Langefors, Borje. Theoretical Analysis of Information
maintenance [2]. Systems (2 vol.), Lund, Sweden, Studentlitteratur, 1970.
[14] Langefors, Borje and Sundgren, Bo. Information Systems
Recursion, as discussed in this article, suggests that Architecture, New York, Petrocelli, 1975.
[15] Machol, Robert E. and Miles, Jr., Ralph F. "The
decoupling is equally valid at the upper levels. The Engineering of Large-scale Systems," in Systems
potential benefits, in fact, may be much greater Concepts: Lectures on Contemporary Approaches to
given that we are trying to minimize the risk of Systems, Ralph F. Miles, Jr. (ed.), New York, Wiley, 1973,
developing wrong, inadequate, or inflexible pp. 33-50.
McKinsey & Comany, Inc. "Unlocking the Computers
systems as we venture into the vague area of [16]
Profit Potential," Computers and Automation, April 1968.
improved decision support for management. [17] Miller, George A. "The Magical Number Seven, Plus or
Decoupling supports an iterative approach for Minus Two: Some Limits on our Capacity for Processing
determining the desired capabilities of the system Information," The Psychological Review, Vol. 63, No. 2
by requiring a formalized, yet tentative, (March 1956), pp. 81-97.
[18] Murdick, Robert G. "MIS Development Procedures,"
specification of objectives at each step of Journal of Systems Management, Vol. 21, No. 12
development for feedback to stimulate user (December 1970).
involvement and to guide further development [19] Myers, Glenford, Jr. Reliable Software Through
efforts. Thus, it helps to prevent design decisions Composite Design, New York, Petrocelli/Charter, 1975.

40 MIS Quarterly / June 1, 1977

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions
Information System Development

[20] Powers, Richard F. and Dickson, Gary W. "MIS


Project Management: Myths, Opinions, and Reality,"
California Management Review, Vol. 15, No. 3 (Spring
1973), pp. 147-156.
[21] Prywes, Noah S. "Automatic Generation of Software
Systems," Data Base, Vol. 6, No. 2 (Summer 1974), pp. 7-
17.
[22] Teichroew, Daniel and Sayani, Hasan. "Automation of
System Building," Datamation, Vol. 17, No. 16(August 15,
1971), pp. 25-30.
[23] Urwick, Lyndall F. "The Managers Span of Control"
Harvard Business Review, Vol. 34, No. 3 (May-June 1956).
[24] Wirth, Niklaus. "Program Development by Stepwise
Refinement," Communications of theA CM, Vol. 14, No. 4
(April 1971), pp. 221-227.
[25] Wirth, Niklaus. "On the Composition of Well-structured
Programs," Computing Surveys, Vol. 6, No. 4 (December
1974), pp. 347-259.

About the Author


Hugh F. Juergens is an Assistant Professor of MIS
in the Quantitative Analysis Department at the
Graduate School of Business, University of
Wisconsin - Madison. He received his B.S., M.S.,
and Ph.D. in MIS from the University of
Minnesota. He previously taught at the State
University of New York at Binghamton and was
formerly an Applications Analyst for Control
Data Corporation. Professor Juergens is currently
engaged in research on Database Design.

MIS Quarterly/ June 1977 41

This content downloaded from 200.21.227.154 on Thu, 01 Oct 2015 03:32:16 UTC
All use subject to JSTOR Terms and Conditions

Вам также может понравиться