On Roles of Models in Information Systems - Intentional Perspectives on Information Systems Engineering - page 22

• The “requirement first” approach recommends that solutions for the comput-

erized parts of an information system should satisfy validated requirements.
The approach also recommends that requirements and solutions should be kept
separated, so that modified solutions can be validated relative to modified
requirements, see [ 46] .

• The “agile” approach recommends that specifications of the computerized parts

of an information system are stated in terms of computer solutions, that is, in
terms of executable specifications, see [ 37] .

The “requirements first” approach is seen as an old-fashioned traditionalist
approach by the “agile” community. The “agile” approach is widely seen by the
“traditionalists” as a “program first - think later” approach which can only be used
for systems of limited size and limited life span.
The classical engineering approach has in particular been criticized for a linear,
heavy handed approach to project management. This is when the classical approach
is compared to competing approaches which are claimed to rely more on iteration
and gradual development of the systems. When subjected to analysis this claim
does not seem to hold [ 2] . Iteration is basic to engineering design, independent of
approach to project management.

3.2 Evolution of Detail

Most solutions to engineering problems are so complex that it is impossible for
anybody to understand every detail at the same time. Information systems of
this complexity are sometimes called unperceivable [ 16] , but are more commonly
called unsurveyable [31] . A basic problem in engineering is to find an approach to
developing solutions to unsurveyable systems.
The most common solution strategy consists in breaking a problem into simpler
subproblems, then to find solutions to each subproblems so that they together form
a solution to the problem. Gabriel Krohn [ 40] developed this divide-and-conquer
strategy into an approach for solving electrical engineering problems [ 41] . He called
the approach The Method of Tearing, or Diakoptics (Greek: dia -through + kopto -
cut, tear). The essentials of diakoptics is the splitting of physical problems into
subproblems that can be formalized and solved independently of each other, then
recombining to give an overall formal solution by taking into account the relations
among the subproblems.
Krohn's method was in the 1960s generalized and applied for information sys-
tems design by Langefors [42] . In the information systems field the approach is
known as The Fundamental Principle of Systems Design [ 16] . A system is seen as
a collection of subsystems with stated external properties, related through a sys-
tem structure which relates the external properties of the subsystems. In order to be
used to its full potential it takes that both the system structure and external proper-
ties of each subsystem can be formally stated, so that the external properties of the
resulting system can be formally calculated. This is possible for some categories of