10.3    Comparison of TEMA with Other Frameworks

Section 10.2 was concerned with evaluation of Tema in itself, both according to internal criteria and in terms of observed use of elements of the framework in three of the studied projects. Section 10.3 follows an alternative strategy in search for confidence in Tema: Comparison with other approaches with roughly equivalent purposes.

10.3.1    Introduction to the comparison

Chapter 5 was concluded by listing three frameworks for enterprise modeling being particularly interesting from the perspective taken on enterprise modeling in this thesis F3, SSM and OOram. Table 10.7 outlines some features of the three frameworks as compared to Tema, using the same criteria as used for presentation in chapter 5.
 
 
Criteria
Approach Background Purpose Type of work Worldview Models Modeling  Fit
F3 ISE I, II I RD Cons Yes Yes v
SSM   I MI RD Cons Yes Yes v
OOram SE I, II MI RD Cons Yes Yes v
Tema   I I D Cons Yes Yes v
Table 10.7: Tema compared with three of the frameworks from chapter 5

The approaches are discussed in more detail below using a condensed tabular form followed by some reflections upon resemblance. Similarities are marked with an equality sign (=) and differences with an inequality sign (#).

10.3.2    From Fuzzy to Formal (F3)

Recalling from chapter 5, F3 is a comprehensive framework for information systems development employing enterprise modeling in the requirements engineering phase. The ambitions of modeling are far more extensive than for Tema, giving rise to both similarities and differences in the two frameworks, see table 10.8.
 
? Framework feature Comparison of framework features
= Preparations to modeling Both F3 and Tema prescribe a preparation phase before enterprise modeling. In F3, this is referred to as preconditions for modeling. Tema prescribe team training.
= Guidelines and driving questions Both F3 and Tema prescribe a more ad-hoc approach to modeling than most other frameworks. F3 offers a number of guidelines and driving questions for their various modeling contexts.
= Enterprise modeling as a team-based activity F3 focuses on modeling as a typical group activity and argue that group based modeling most often is beneficial as compared to individual modeling. Advantages are more ideas, ownership and responsibility to the final results, and increased knowledge in the organization. Tema is designed as a group based approach in all respects.
= Forums for modeling F3 advocates the use of modeling seminars, being in effect equivalent to meetings in Tema.
# Purposes of modeling F3 is more ambitious regarding the purpose of modeling than Tema. This has influence on the need for formal modeling languages and need for computer support in modeling sessions.
# Modeling language F3 prescribes an extensive ontology of what Tema refers to as the Universe of Production. This is quite in opposition to what Tema does.
# Subsidiary and intermediate artifacts F3 does not seem to discuss what Tema refers to as subsidiary or intermediate artifacts. Annotations and fuzzy concepts may play the role of subsidiary artifacts, but not to the extent discussed in Tema.
# Development versus dissemination F3 does not discuss the dissemination phase of enterprise modeling in any notable detail. Models are neither assessed to be different in the two phases, whereas Tema stresses the different qualities.
# Medium for modeling The use of media for development and dissemination is very different in F3 and Tema. F3 relies on the digital medium for all purposes, while Tema prescribes the use of static media like paper in development.
# Size of framework and models Both models and the documented framework are very comprehensive in F3 when compared to Tema
Table 10.8: A comparison of F3 and Tema

General considerations of F3

F3 incorporates a number of relevant ideas, particularly regarding the modeling process. Their guidelines and driving questions may be of value to Tema modelers, too. The F3 guidelines are claimed to have been developed from extensive practical experience with modeling.

The focus of F3 is clearly more stable, long-term human activity systems than Tema. F3 advocates representation of all available information (an objective of F3 is not to loose any information in requirements acquisition), seeing enterprise models as repositories of knowledge (Bubenko and Kirikova, 1994:9, italics added):

"Enterprise Modelling is seen as a building of repository of the knowledge to be used to support to carry out the requirements acquisition process, "
Tema is more dedicated to enterprise modeling as sense-making and enterprise models as a means to aid in the sense-making process.

10.3.3    Soft Systems Methodology (SSM)

SSM is a general methodology for inquiry into problem situations, with actors trying to make sense of the reality they encounter. The focus on human sense-making and to some degree similar activities make SSM and Tema resemble each other. Still, table 10.9 reveals some substantial differences.
 
? Framework 
feature
Comparison of framework features
= Purposes of modeling Both SSM and Tema focus on human sense-making as the overall purpose of applying the approach.
= Intermediate and subsidiary artifacts SSM proposes the use of root definitions, metaphors and rich pictures in the sense-making process. This coincides with Tema's notion of both subsidiary and intermediate artifacts.
= Systematic versus systemic SSM claims to be a systemic approach, not systematic. The belief in an ad-hoc approach to enterprise modeling is common to both SSM and Tema.
= Team-based approach The social aspect of human sense-making is central to both SSM and Tema.
# Aspects  The notion of streams in SSM represents a more varied analysis than Tema offers in its current version. The cultural stream has no corresponding component in Tema (although it may easily be incorporated, as the Universe of Production is not restricted).
# Preparations to modeling The preparatory stage in the development phase of Tema has no equivalent in SSM.
# Development versus dissemination SSM does not differentiate between development and dissemination phases the way Tema does.
# Use of media and forums Although both SSM and Tema rely on use of static media like paper, and meetings as the main forum, Tema goes a step further by prescribing alternative media and forums for the dissemination phase.
# Maturity SSM is a far more mature and "proven" approach than Tema, as it is claimed to have been under development and in use for over 20 years. Experiences from practical use and adaptations of SSM for particular applications are also reported (e.g., Checkland and Scholes, 1990).
Table 10.9: A comparison of SSM and Tema

General considerations of SSM

SSM is a methodology according to Checkland and Scholes (1990:299) a set of principles of method. It must be tailored to each specific problem situation to be useful.

Tema is developed particularly for enterprise modeling of work associated with design risk, corresponding to SSM's notion of ill-defined and tacky problem situations. It includes concrete proposals for techniques and tools. Still, Tema must also be adapted before practical use, and activities in SSM may be tailored to fit the development phase of Tema.

10.3.4    Object-Oriented role analysis and modeling (OOram)

OOram is an object-oriented software engineering method. One of the activities in applying the method is enterprise modeling for the purpose of human understanding or for simulation. Table 10.10 summarizes the comparison with Tema.
 
? Framework feature Comparison of framework features
= Subsidiary artifacts OOram provides a number of different views that all describe aspects of the enterprise model. The views may be considered to correspond to Tema's subsidiary artifacts.
= Lack of strict method Although OOram presents an overall method, it is stressed that modeling necessarily have to be opportunistic. This is in line with the pool of techniques proposed in Tema.
# Purposes of modeling OOram aims at being an approach for both human understanding and computer simulation. This is in contrast to Tema, intended for sense-making only.
# Modeling language The OOram language is formally defined in advance, being a restriction of the Universe of Production not found in Tema.
# Preparations to modeling OOram does not prescribe an equivalent to team training in Tema
# Development versus dissemination OOram does not discuss what Tema refers to as the model dissemination phase. There are some discussion of how to approach and involve actors in the modeling process, but not to the degree found in Tema.
# Team-based modeling While Tema has been developed for modeling in teams, OOram does not seem to stress the social aspect of modeling.
# Forums and media OOram does neither discuss alternative media or forums for enterprise modeling. 
# Maturity OOram is relatively mature, as the ideas are claimed to have been under development and in use for about 20 years (Reenskaug, 1996:xix).
Table 10.10: A comparison of OOram and Tema

General considerations of OOram

OOram has much in common with F3 as it is a method for software development. The ultimate purpose of modeling is not the understanding, but to be able to construct a piece of software. The consequence is different stress put on activities in modeling.

Another factor that causes differences is that OOram restricts the Universe of Production through its reliance on the role modeling language and the number of views, while Tema is more open (although the proposed pool of techniques also constrain the UoP to some degree).

10.3.5    Concluding remarks on the comparison

The main purpose of the comparison has been to increase the confidence in Tema. Underlying this purpose is the belief that if a proposed feature of Tema is found in another approach as well, it seems more plausible that it contributes positively to enterprise modeling. If a feature is not found, there ought to be a difference in objectives of the two frameworks or a difference in underlying beliefs.

Scanning through the tables in the current section, most of the features of Tema are found in one or more of the other frameworks. E.g., the idea of team training can be found in F3, as they require a set of preconditions to be satisfied before modeling starts (including requirements to domain knowledge and actor competence). The idea of a flexible modeling approach is found in all three approaches, although not in the form of a pool of techniques.

Tema's focus on different media and different forums in the development and the dissemination phase is not found in the other approaches, at least not to the same degree. SSM advocates the use of static media like paper, but does not discuss dissemination phase in any significant detail.

To conclude, the comparison have illustrated that a number of features of Tema can be found in other approaches as well, presumably contributing to confidence in Tema.