Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide


  1. 1. Philosophical Foundations for A Unified Enterprise Modelling Language Gerald R. Khoury Institute for Information and Communication Technologies University of Technology Sydney PO Box 123 Broadway NSW 2007 Australia Simeon J. Simoff Institute for Information and Communication Technologies University of Technology Sydney PO Box 123 Broadway NSW 2007 Australia Abstract information environments become more In this paper, we lay the groundwork for an complex, more sophisticated methods are understanding of enterprise systems modelling needed to manage these environments that is couched within the framework of effectively. EA’s help manage this change and philosophic, linguistic and cognitive theory. overcome the problems of building isolated IT This understanding is used as the basis for a solutions that fail to support an enterprise’s novel approach to developing unified enterprise vision, goals and objectives. “Understanding modelling languages. and visualising complex businesses enables you to identify and address areas that might be A case is presented for viewing enterprise constraining business performance. Enterprise models as metaphors. It is shown how these modelling helps you focus on those areas you metaphors can be structured as part of a can change, how these areas are currently hierarchy, in keeping with Way’s dynamic type functioning, how they might be optimised, and hierarchy theory (Way, 1991). These how any changes might impact other areas.” observations are extrapolated to show how (Fraser and Tate, 1995) highly conceptual metaphors can be used to describe all enterprises. Finally, there is a However, there are barriers to achieving EA’s discussion on the codification and development full potential. Traditional EA models are often of these metaphors as the basis for a unified unwieldy and difficult to navigate and explore. enterprise modelling language that could be The problem is exasperated by the fact that the used to model all enterprise domains. stakeholders to which these architectures are directed have “varied backgrounds, and Keywords: Enterprise – Architecture – Modelling – technical and non-technical skill sets and Conceptual – Systems – Metaphor interests.” (Armour et al., 2003) Typically, 1 Introduction these stakeholders include the CIO, a wide range of business users, system users, IT While enterprise architecture (EA) is still in its developers, systems analysts and systems infancy as a research area, this has not architects. In fact, Bemelman and Dennis’s prevented it from becoming an important and investigation of architecture from a users' point firmly established discipline within the IT of view (in (Rostad, 2000, p.136)), concludes industry. In fact, EA’s are growing in that "the inherent levels of complexity and importance as tools for managing change within detail in most of the current architectures will today’s highly dynamic, demand driven and become a major impediment to acceptance of highly competitive business environments. As these architectures in the industry." the rate of technological change increases and
  2. 2. 2 The Difficulty with Enterprise Modelling This leads to an EA model that provides a The ability to create effective EA’s rests on the fragmented view of the enterprise with poor ability to create “simple models that are easily explanatory power and little flexibility for communicated …" (Rostad, 2000). future planning. Unfortunately, current enterprise models are There have been serious attempts to develop a complex, difficult to produce and hard to single language that can be used to describe the understand because they can only be produced entire EA domain. Unfortunately, these using variety of unrelated modelling languages. attempts have suffered from significant While specialised languages have been drawbacks that have prevented them from developed to model specific domains such as achieving industry acceptance. One of the most application, infrastructure or network notable proposals is the ‘conceptual graph’ architectures, none of these languages support notation. (Zachman and Sowa, 1992) Zachman the creation of high-level, conceptual systems and Sowa proposed the use of conceptual models that extend across the enterprise. "The graphs, a form of symbolic logic, to provide a range of phenomena addressed by enterprise natural language that can be used as a unifying modelling stretches multiple disciplines, and ontology for all the models required by the accordingly many modelling languages and Zachman framework. practices are used. There is no single person/profession who would be able to However, Zachman and Sowa first suggested guarantee the consistency of all models the application of conceptual graphs to produced …" (Bernus et al., 1996) Yet, enterprise modelling more than a decade ago. successful systems implementations depend on Yet, there is scant evidence of the application of the availability of integrated models that are conceptual graphs to real world EA problems. consistent and non-contradictory. (Roussev and The primary reason for this is probably that Rousseva, 2004) conceptual graphs are too complex for general use as an effective modelling tool. This is In fact, there are currently no standards to guide evidenced by the work of Polovina (1993), who the development of EA models: "Today, no investigated the suitability of conceptual graphs single existing modelling language by itself is as a way of modelling strategic management capable of modelling all necessary aspects of an problems within an accounting domain. enterprise." (Noran, 2003) This leads to several Polovina found that "However, despite their problems. Firstly, enterprise architects are strong prima facie attractiveness … the inherent required to have expertise in multiple modelling complexity of conceptual graphs fundamentally languages in order to create enterprise models. undermined them as a viable tool, other than for Secondly, the lack of consistent semantics and a very trivial problems well below the level weak ontological foundation makes it difficult needed to be viable for strategic management to create coherent and consistent enterprise accountancy." Although Polovina was working models. Lastly, the resulting models are beyond in a specialised domain, there are many the comprehension of those who are not similarities between the modelling he was modelling experts and yet are interested in EA. testing and the modelling of EA’s. In addition, This includes technology and business leaders, the senior practicing accountants who were his perhaps the most important users of EA’s. experimental subjects have similar analytical Ultimately, the use of multiple EA languages competence to enterprise architects. that cannot be effectively integrated, prevents Other attempts to achieve an integrated the achievement of the prime goal of an EA: the enterprise modelling language include the development of an integrated view of the Hatley/Pirbhai (H/P), Quantitative QFD (Q2FD) enterprise. (Maier and Rechtin, 2000) In and WfMC Meta-model approaches. However, particular, EA’s that are built using component- all of these methods are highly restricted in based frameworks (such as the popular terms of domain. As a result, they are unable to Zachman framework) are fundamentally flawed deliver effectively the benefits that would be because they model the enterprise as a set of forthcoming from a method that integrated well independent structures with discrete boundaries. over the entire EA domain. Also, "even in these
  3. 3. domains the models are not in very wide use 4. metaphor cannot be reduced to any literal …" (Maier and Rechtin, 2000, p.219) statements of comparison, and; 5. metaphor can actually create similarity between It should also be noted that contemporary EA previously dissimilar ideas. approaches have been developed informally. (Way, 1991, p.48) This suggests that a scientifically grounded, theoretically sound approach to EA Way acknowledges that there are criticisms of development may offer improvements over the interaction view, however, she believes that contemporary EA methods and frameworks. despite these “the interaction view has been the In this paper, a new theory is expounded that most widely accepted and influential view of can be used to develop an integrated approach metaphor.” (Way, 1991, p.50) Subsequently, to EA modelling. This theory makes use of an Way’s new theory, the DTH theory, is understanding of metaphor and their relation to developed as a variation and improvement on system model hierarchies. The following the interaction view theory. The DTH theory is section provides an understanding of the place thus purported to explain the results of of metaphor within type hierarchies. We then empirical tests where all other theories fail. look at the connection between models and The DTH theory uses standard concepts from metaphors and this provides a link between type graph theory: “A type hierarchy is a network of hierarchies, metaphor and model hierarchies. concepts which are organized according to levels of generality. … So the links connecting 3 Dynamic Type Hierarchies the supertypes and subtypes of the semantic The dynamic type hierarchy (DTH) theory network represent going from an instance of a “involves a theory of metaphor that incorporates supertype to a more specific instance of that Sowa’s conceptual graphs, dynamic type supertype. … an instance of a subtype entails hierarchies, and Max Black’s interaction that it is also an instance of the corresponding approach.” (Way, 1991, p.125) Numerous supertype.” (Way, 1991, p.21) The DTH is a theories have been developed explaining how generalisation hierarchy where the entities are metaphors are used and why they work. These categorised based on the common attributes. In theories include emotive theories, the a generalisation hierarchy, the higher-level class substitution approach, the comparison theory, (supertype) share the common attributes of the metaphor as analogy, the controversion theory lower level class (subtype), while subtypes and the anomaly theory. Way surveys and inherit all the properties of its supertype. analyses these approaches and illustrates the shortcomings and contradictions that beset “The DTH theory holds that the similarity them. Way finds that the interaction view, while found between the tenor and vehicle of a not perfect, is the most promising contemporary metaphor1 is not an intersection of their theory of metaphor and that “much of the properties; rather, it is generated by finding a experimental data of metaphor is either common and more abstract supertype that the compatible with or actually supports aspects of two share.” (Way, 1991, p.40) The supertype is the interaction view.” (Way, 1991, p.124) a generalisation of the attributes of the connected subtypes. “In metaphor, what is Max Black originated the interaction view common between the vehicle, and tenor is not (Black, 1962). Way summarises the main points an intersection of a list of features at the level of of the interaction view of metaphor as follows: the tenor and vehicle, but a supertype, which is 1. metaphor involves entire systems of assumptions higher up on the semantic hierarchy and under and ‘commonplaces’ which are associated with which aspects of both the vehicle and tenor the terms involved; domains fall. Furthermore, which supertypes 2. that the metaphorical process works like a filter, are chosen, assuming that there are several in with the associated ideas of the secondary common, is a function of the context and the subject (vehicle) hiding, highlighting and direction of the attribution of the metaphor, that organizing aspects of the primary subject; 3. understanding metaphor often involves a shift in 1 Here, the terms tenor and vehicle can be reasonably meaning; substituted for target and source respectively.
  4. 4. is, the metaphor is attributing features from the topic or tenor, and the source is often called the vehicle domain by abstracting them to a secondary system or the vehicle. common supertype and then using that Contemporary research has shown us that supertype to pick out the corresponding features metaphor, far from being just a figure of speech, of the tenor.” (Way, 1991, p.129) is central to everyday communication and According to the DTH theory, metaphor learning. (Lakoff and Johnson, 1980) (Lakoff operates through a higher-level supertype that and Nunez, 2000) In fact, it has been argued connects the tenor and vehicle types. If this that only by relating a new concept (the ‘target’) supertype is not found to already exist in the to a well understood, everyday object (the type hierarchy, it is dynamically created, just as ‘source’), can one develop new knowledge and in some cases "metaphor creates the similarity" understanding and that without this process, it is rather than describing some existing similarity. impossible for humans to carry out abstract (Black, 1962) Two separate clusters of thinking. schemata thus become associated through a As our understanding of metaphor has supertype and this allows a migration of developed, the links between models and concepts from the vehicle to the tenor. This metaphor have become more apparent. This is exemplifies what Black terms “systems of most evident in the case of linguistic models. associated commonplaces” (Black, 1962): the Black was one of the first to make the case for complex and far reaching web of associations the connection between models and metaphor in and meaning that imbue every metaphor. language and philosophy (1962), and in later Way provides as an example the metaphor ‘the years, Black becomes further impressed "by the car is thirsty’. This metaphor “involves a tight connections between the notions of models violation of a constraint, in that thirsty is an and metaphors. … Every metaphor is the tip of attribute of an animal, not a vehicle. The new a submerged model." (1979, p.31) More supertype, Mobile-entities that require liquid, recently however, these early notions on the can have both tenor and vehicle fall under it connection between models and metaphors have without violating any semantic constraints.” been extended. Lakoff, for example, (1991, p.130) "Note that the set of properties of demonstrates that mathematical models are also the supertype will be smaller than those of its metaphors. (Lakoff and Nunez, 2000) subtypes; in this way the supertypes are able to According to Roussev and Rousseva, the act as filters on lower level concepts. … The mechanics of modelling are the same as supertypes VEHICLE picks out one aspect of a metaphor, and modelling evokes the same car, while another supertype, say STATUS- cognitive processes as metaphor. Indeed, SYMBOL, would pick out or filter different “Models are always grounded in a kernel aspects." (Way, 1991, p.161) metaphor, or metaphors, and hence stimulate many of the same cognitive processes.” 4 Models and Metaphors (Roussev and Rousseva, 2004) Lakoff and Johnson define metaphor as: “… Interestingly, the definition of a model - “A understanding and experiencing one kind of model is a representation of something else …” thing in terms of another.” (1980) Their ISO/ANSI quoted in (Szegheo, 2000), is ontology parallels the description of metaphor remarkably similar to the definition of a used widely within contemporary studies of metaphor as an “image that represents one thing cognitive linguistics, where the terms source as something else in order to explain it better and target refer to the conceptual spaces …”, (Johnson, 1994). In fact, an empirical connected by the metaphor. The structure of the argument can be made for the assertion that source domain is projected onto the target metaphors are models. According to Herbert domain in a way that is consistent with the Stachowiak (as referenced in (Ludewig, 2003)) inherent target domain structure (Lakoff, 1993). an artefact must satisfy the following three Elsewhere in the research literature, the target is criteria in order for it to be identified as a variously referred to as the primary system, model:
  5. 5.  Mapping criterion: there is an original object or (Black, 1962) has been set up, because these phenomenon that is mapped to the model. labels are not simply literal: they are metaphors which give rise to supertypes which  Reduction criterion: not all the properties of the encapsulates much of the understanding of what original are mapped on to the model, but the the system is and how it operates and so an “entire web of associations and implications” model is somehow reduced. On the other hand, (Way, 1991, p.36) has been altered. the model must mirror at least some properties of Therefore, the labels that are assigned to the original. 2 enterprise systems, can themselves, be used as  Pragmatic criterion: the model can replace the models if they are terms that have associated meanings outside of the system context. original for some purpose, i.e. the model is useful. In fact, the metaphorical approach to It will be recalled that a metaphor connects a organisational modelling is founded on the source domain to a target domain. This satisfies observation that the reality of organisational the first criterion. When metaphor is used, only structures are not concrete, but abstract, certain characteristics of the source are mapped indefinite, and perhaps even undefinable, to the target, and this depends upon the context entities. There is no objective reality of an in which the metaphor is being used. (Way, organisation. The significant structures within 1991) Thus, the second criterion is satisfied. an organisation are actually social constructs, Finally, we use metaphor in order to explain and as such, they are imbued with cultural and something better (Johnson, 1994) especially social meaning, presenting differently to every when that something is new and novel. person according to their immediate concerns. (Ludewig, 2003) Therefore, metaphor satisfies In modelling organisations we are not simply all three of Stachowiak’s criteria defining a developing abstractions of an objective and model, and metaphor is a type of model. More concrete structure. Rather, we are saying 'this is specifically, metaphor is a type of conceptual how we wish the organisation to be understood’. model. (Allen, 1997) For example, ‘we want our enterprise to be seen Furthermore, it can be observed that the labels as developing relationships with customers, not given to enterprise systems are actually exploiting them’. As Black says, the metaphor metaphors (and by induction, models). Take as actually helps constitute aspects of reality an example, an enterprise system labelled (Black, 1962) and we create meaning through 'Customer Relationship Management System'. If the development of metaphors for these this were simply a label, and not a metaphor, structures. then there would be no significant effect in changing this label to 'Customer Exploitation 5 Model Hierarchies Management System' as long as there were no There are two conditions for the development of changes made to the actual systems to which an effective unified model. Firstly, the unified this model refers. But in fact, this label change model must be capable of representing the fundamentally changes the perceptions that will semantics of the various sub-domains that are be generated around this system. A completely being modelled, albeit, at a higher level of different set of “associated commonplaces” abstraction. Secondly, the unified model must present a strong ontology so that the model can 2 It must be noted, that this is contrary to some popular be interpreted unambiguously. A single, well- notions of models. For instance, Ghyczy suggests that defined and formalised modelling language will models “exhibit a one-to-one correspondance” with their provide a strong ontology. Unfortunately, no source and that you can “transfer everything you know current modelling languages can cover the about the source domain into the target domain” if you have a good model (Ghyczy, T. v. (2003) The Fruitful semantic breadth needed to describe the Flaws of Strategy Metaphors Harvard Business Review, different types of enterprise systems, without 86-94.) Clearly, this is false: models are only useful inordinate complexity. In this section we deal because they are abstractions of the source domain, and with the problem of developing a semantically by implication, this means that everything must not be unified enterprise model. In the following transferred from the source to target domains.
  6. 6. section, this model is used to develop an this metaphor could provide a useful ontology ontology that can be applied to formally and a desirable set of associated meanings and describe any sub-domain. implications. However, if ‘Enterprise A’ were a finance company, then this metaphor would not It has already been shown that metaphor can be be useful or desirable. viewed as part of a dynamic type hierarchy. For the purposes of this argument, the types in this On the other hand, if we created a metaphor that hierarchy will be restricted to the domain of compared a supertype of an enterprise to enterprise models. Note that, since it has been another type on the same level as that shown that metaphors are a special type of supertype, then the metaphor would be relevant model, this restricted domain still includes to all enterprises. In fact, it would provide a metaphor. unified concept that could be applied to any of its sub-models. This is illustrated in Figure 2. Within any enterprise, there is a set of models The task remains, to identify, or invent, a representing various levels of abstraction. At supertype of all enterprise systems models. the highest level is the model of the enterprise itself. This may, or may not be formalised. At lower levels of abstraction are more detailed Enterprise A models such as system and function models. It can be observed that larger scale structures tend to be modelled using more general, 'conceptual' models and metaphors while smaller scale structures tend to be modelled using more System A System B specific, 'concrete' models and metaphors. Subsequently, enterprise system models can be formed into a generalisation hierarchy, where the higher-level classes (supertypes) share the Function A Function B Function C Function D common attributes of the lower level classes (subtypes), while subtypes inherit all the Figure 1 - Enterprise System Hieararchy with properties of its supertypes. Enterprise as Global Supertype Figure 1 shows an enterprise system hierarchy where ‘Enterprise A’ is the universal supertype. According to Way, "metaphor cannot take place Enterprise Supertype by comparing the properties of independent systems on the same level", i.e. at the same level of abstraction. "In order to properly make Enterprise A Enterprise B the comparison, we must search for higher order concepts. Comparison of properties on one level only make sense as a comparison with respect System A1 System A2 System B1 System B2 to a common, more abstract, higher level property." (1991, p.144) Therefore, it is possible to create a metaphor between Function A1 Function A2 Function A3 Function B1 Function B2 Function B3 ‘Enterprise A’ and another type at the same level of abstraction as Enterprise A that could Figure 2 - Enterprise System Hieararchy with New be used to describe features of the enterprise Enterprise Supertype and each of its component structures (‘System A’, ‘System B’, ‘Function A’, ‘Function B’ etc) There are any number of abstract concepts that by generating a higher level type. However, as could be used as a metaphor source for every enterprise is unique, this metaphor may enterprise systems. These include the not be relevant to other enterprises or their description of organisations as machines, component structures. For example, we might organisms, brains, cultures, political systems say, “Enterprise A is transportation”. If and learning organisations. (Morgan, 1996) ‘Enterprise A’ was, say, an airline company, Note however, that some of these metaphors use
  7. 7. source objects that are at the same, or lower metaphor3 appear to be strong, especially as any levels of abstraction than the enterprise type. society based concepts and nomenclature are Therefore, they are not suitable as a unifying likely to be very familiar to all users of the metaphor for all enterprises. Take as an enterprise architecture. While there are other example, the description of an organisation as a metaphors that would satisfy the semantic machine. This may describe certain aspects of requirements of a unifying supertype, the an organisation, however it is a highly concrete concept of society is perhaps a more promising metaphor and is too confining to describe all metaphor for this problem, since it has a aspects of all systems. For instance, it cannot particular relevance and efficacy for enterprise describe the ability of an organisation to learn, systems modelling. regenerate or develop purpose. The metaphor of Formalisation and codification of the enterprise ‘the learning organisation’ is however more model supertype will provide a foundation for a conceptual and flexible, explaining perhaps it modelling language that can be used to describe greater popularity as an enterprise metaphor. component structures of any enterprise, but to Another highly conceptual metaphor is ‘the do this we need to understand what a model of enterprise as a societal structure’. In the society looks like. Fortunately, one such model following section, this metaphor is adopted and has been developed by Giddens (Giddens, developed as a basis for a unified systems 1984). According to Giddens' Theory of language. Structuration, there is interdependency between humans (actors) and societal structures 6 A Societal Language for Unified Systems (resources and rules) that is manifest through Modelling specific actions. Perhaps one of the conceptual structures most familiar to humans is that of human society. The notion of an actor is extended here to Society is a larger scale structure than an include, not only individuals, but also any agent enterprise (societies contain enterprises, but that can exert power in order to produce an enterprises aren’t normally viewed as effect. To this end, the terms actor and agent are containing societies), and accordingly, it can be used interchangeably. Resources are “structured used as a supertype of all other types in the properties of social systems, drawn upon and enterprise system type hierarchy. That is, a reproduced by knowledgeable agents in the society-sourced metaphor will be at a higher course of interaction.” Resources are of two level of abstraction than any of the models that types. Allocative resources are material describe an enterprise or its components. resources involved in the generation of power Therefore, the concept of society has the and derive from human dominion over nature. semantic breadth to describe all enterprise Authoritative resources are non-material and systems. derive from the capability of harnessing the activities of human beings. (Walsham and Han, Is a society-sourced metaphor a good choice for 1991, p.84) Rules refer to the sanctioned modes modelling enterprise systems? It is clear that all of conduct, and an action is an activity that is enterprises are social phenomena and that performed. "Structures, as 'rules and resources', computer systems are socially embedded do not do anything, but they have their effect phenomena. In lieu of a social context, through being known and used by actors." computers can have no meaning or value. (Parker, 2000) (Winograd and Flores, 1987, p.78) It is also widely noted that many systems failures occur, This provides an ontology consisting of actors, not because of technical issues, but because the resources, rules and actions that can be used as relevant socio-political concerns were not the basis for developing a formal modelling evaluated and addressed. A society-sourced metaphor is likely to be stronger than many 3 A measure for the degree to which models expressed other metaphors for modelling these aspects of using this metaphor will succeed in being interpreted by the audience as intended by the creator. Biemans, F. P. the enterprise. Thus the ‘pragmatics’ of this M., Lankhorst, M. M., Teeuw, W. B. and Wetering, R. G. v. d. (2001) Dealing with the Complexity of Business Systems Architecting Systems Engineering, 4, 118-133.
  8. 8. language. These four structures might be used used to replace a myriad of disparate, and as nodes, for instance, in a graphical language possibly, contradictory metaphors, the resulting that connects these nodes using arcs according models have the following qualities: to well-defined rules. Furthermore, the nodes  Greater explanatory power (a consistent themselves could be used as the minimal metaphor/language is used to model the entire common supertypes of four different type hierarchies, where the subtypes within each of organisation which allows the user to develop an these hierarchies inherit the attributes of its effective mental map of the organisation). parent. This would provide the language with a high level of semantic power, despite the fact  Greater flexibility for the management of change that only four normative types exist. and strategic planning (current disjunctions between systems, business units etc. are not ‘hard 7 Conclusion wired’ into the model). "Enterprise models are useful only if they are used. They will be accepted by users  Avoiding loss of information that might occur in as a tool if they are simple to understand, translating from one architectural view to another easy to use, computer-supported, and if they provide a realistic image of the and ensure cross-view consistency (Armour et al., reality. This explains the failure of many 2003) approaches proposed in the past …" (Solberg, 2000)  Explicit capture and representation of business needs within the EA models (Armour et al., 2003) Without a unified enterprise modelling language, there is no way to develop coherent  Reducing cognitive load placed on a user that and consistent enterprise wide models of IT needs to understand a complex set of architectural systems. Thus, the development of a unified EA views (Armour et al., 2003) modelling language is a highly coveted goal. Its achievement would overcome some serious Research is continuing in this area and includes shortcomings with current EA approaches and the development of a tool to support the improve the efficiency and effectiveness of application of these languages, as well as the enterprise architects and systems planners. evaluation of unified languages that are Unfortunately, the development of such a generated using the theory described in this language has remained elusive. paper for ‘real-world’ enterprise architecture A philosophy-based approach to the problem of problems. unified EA modelling has the potential to solve a problem that has proven to be intractable 8 References using more traditional and direct approaches. Allen, R. B. (1997) In Handbook of Human-Computer By applying an understanding of type Interaction(Eds, Helander, M., Landauer, T. K. hierarchies to the selection of system and Prabhu, P.) Elsevier Science. metaphors, it is possible to identify a metaphor Armour, F. J., Kaisler, S. H., Getter, J. and Pippin, D. that provides a scope that is wide enough to 2003 A UML-driven enterprise architecture case study System Sciences, 2003. Proceedings of the cover all of the functional elements of the 36th Annual Hawaii International Conference enterprise’s component structures. After all, it is on Hawaii better to have a single metaphor that covers the Bernus, P., Nemes, L. and Morris, R. (1996) In entire domain, for “when discourse becomes Modelling and Methodologies for Enterprise full of conflicting metaphors, it may be difficult Integration(Eds, Bernus, P. and Nemes, L.) for the uninitiated to keep their bearings.” Chapman and Hall, London. (Johnson, 1994) Biemans, F. P. M., Lankhorst, M. M., Teeuw, W. B. and Wetering, R. G. v. d. (2001) Dealing with the This metaphor can then be used to develop a Complexity of Business Systems Architecting unified language that can model all of the Systems Engineering, 4, 118-133. enterprise’s systems. Since a single metaphor is
  9. 9. Black, M. (1962) Models and Metaphors: Studies in Walsham, G. and Han, C.-K. (1991) Structuration Theory Language and Philosophy, Cornell University and Information Systems Research Journal of Press, London. Applied Systems Analysis, 17, 77-85. Black, M. (1979) In Metaphor and Thought(Ed, Ortony, Way, E. C. (1991) Knowledge Representation and A.) Cornell University Press, London, pp. 19-43. Metaphor, Kluwer Academic Publishers, Fraser, J. and Tate, A. 1995 The Enterprise Tools Set - Dordrecht. An Open Enterprise Architecture Proceedings of Winograd, T. and Flores, F. (1987) Understanding Workshop on Intelligent Manufacturing Systems Computers and Cognition: A New Foundation (IMS), International Joint Conference on for Design, Addison-Wesley, Reading, Artificial Intelligence (IJCAI-95) Massachusetts. Ghyczy, T. v. (2003) The Fruitful Flaws of Strategy Zachman, J. A. and Sowa, J. (1992) Extending and Metaphors Harvard Business Review, 86-94. formalizing the framework for information Giddens, A. (1984) The Constitution of Society: Outline systems architecture IBM Systems Journal, 31, of the Theory of Structuration, Polity Press, 590-616. Cambridge. Johnson, G. J. (1994) Of metaphor and the difficulty of computer discourse Communications of the ACM, 37, 97-102. Lakoff, G. (1993) In Metaphor and Thought(Ed, Ortony, A.) Cambridge University Press, Cambridge, pp. 202-251. Lakoff, G. and Johnson, M. (1980) Metaphors we live by, University of Chicago Press, Chicago. Lakoff, G. and Nunez, R. E. (2000) Where Mathematics Comes From, Basic Books, New York. Ludewig, J. (2003) Models in software engineering - an introduction Software and Systems Modelling, 2, 5-14. Maier, M. W. and Rechtin, E. (2000) The Art of Systems Architecting, CRC Press, Boca Raton. Morgan, G. (1996) Images of Organization, SAGE Publications. Noran, O. (2003) An analysis of the Zachman framework for enterprise architecture from the GERAM perspective Annual Reviews in Control, 27, 163-183. Parker, J. (2000) Structuration, Open University Press, Buckingham. Polovina, S. (1993) Loughborough University of Technology. Rostad, C. C. (2000) In Enterprise Modeling: Improving Global Industrial Competitiveness(Eds, Rostadas, A. and Andersen, B.) Kluwer Academic Publishers, Boston, pp. 119-136. Roussev, B. and Rousseva, Y. (2004) Software Development: Informing Sciences Perspective Issues in Informing Science and Information Technology, 1, 237-245. Solberg, C. (2000) In Enterprise Modeling: Improving Global Industrial Competitiveness(Eds, Rostadas, A. and Andersen, B.) Kluwer Academic Publishers, Boston, pp. 183 -199. Szegheo, O. (2000) In Enterprise Modeling: Improving Global Industrial Competitiveness(Eds, Rostadas, A. and Andersen, B.) Kluwer Academic Publishers, Boston, pp. 21-32.