weight (which implies a better measure of objectivity). This
can be done using the Delphi Technique, where evaluation
results can be tabulated. The finally accepted weights for
the phases and functions can then be recorded in
appropriate columns of the Evaluation Matrix (see Appendix
1).
$ STEP 3: Here the contingency model for the evaluation
of the different functions is defined. This model would
depend on the nature of the problem area for the object
system, one of which could be the aspects (or
perspectives) and level of decomposition relevant for the
functions, and provided for by the methodologies. The
aspects can be process, information, behaviour, and
change; and the level of decomposition may be OSAD
(Object System Analysis and Design)and /SAD (Information
System Analysis and Design) [Slooten, 1992]. In considering
OSAD and ISAD, a simple table as shown in Figure 5 can
be used to ‘fix’ criteria per activity.
STATIC & ENTITY TYPE
specific. Decisions can depend on far reaching concerns as.
management style of the organisation, organisational size;
and these issues may not be mutually exclusive. Perhaps
decision making "elements" within the Object System may
consider Collateral Analysis as a means of articulating the
problems at this stage [Stamper and Kolkman, 1991]. One
Organisation's "costly" project may be cheap for another,
depending on whether cost is viewed from the short or long
term expectations. A cumbersome methodology for staff of
one Organisation may be fair or easy for another
depending on staff experience and aptitude. Decision
variables can also be legal because apparently at each
stage of the system development process,legal implications
tend to exist. The parallel development of Laws and their
related information systems is advocated [Stamper, 1992]
for this reason.
DYNAMIC & EVENT
TYPE
DESIGN cn
P
PHASE = = E = =
= =
= = © ea
FUNCTIONS ea = = 2
(ACTIVITIES) SE = = e El =
OSAD «e
ANALYSIS
ISAD a
OSAD
ISAT
OSAD
ISAD
Figure 5: Pinpointing Criteria using a Framework with two Decomposition levels
As an example, the criteria for evaluating the activity
(function) "Analysis" within the Design phase is considered.
This activity within the problem domain of our Object
System, prepares the conceptual and functional models of
the database system and so exists at both levels of
decomposition. At the OSAD level, it is more oriented to
Organisational change and at ISAD level, more oriented to
information processes. The boxes are appropriately marked
as shown. Therefore the criteria for evaluating a
methodology against the activity "Analysis" could be:
C1 = OSAD / Change and C2 = ISAD / Process.
+ _STEPS 4 AND 5: These mark a decision process
tempered by the global issues inherent in the Object
System. These could be Institutional, Economical and
Technical and are usually Organisation and System
+ STEPS 6 & 7: À value can be assigned to the criteria in
step 4. At this point, step 5 results introduce the filtered
methodologies into the evaluation model in a matching
process. For methodology M, criteria C1 and C2 are
evaluated (again subjectively per evaluator ) based on how
M provides for C1 & C2 in terms of origin and experience,
development process, model, iteration and tests,
representation means, user orientation, tools and
prospects,and generally how they satisfy the expectations
of a good Methodology. This results in an unweighed rating
of methodologies for that activity.
+ STEP 8: Here the ratings are weighted, first per function,
then per phase, and then a simple sum of weights is
determined for all phases. This would conclude the
preliminary stage of the process of selection per
264
methodo
(OSAD,!
criteria).
+ STEP
soll
(per eval
level of
techniqu
Analysis
in a tab
methodo
consider
whole, b!
be seen
or otherv
within th
embeddt
+ STEP:
Strategy
(i) Protot
or (iii) A
The dete
degrees
relevant
Strategy
acceptat
Harper,
To arrive
stated cc
in order t
of the de
CONCL
The Sel
any Info
understz
Develop
a good r
necessa
