This discussion paper refers to the article "Matters on (Meta-)Modeling" of Thomas Kühne. Many of his concepts and proposals are appreciated, as e.g. the clear distinction of type and token models or of prominent relationships like instantiation, generalisation or "meta-ness". However, for some of the presented views and definitions alternative approaches may be followed. This concerns e.g. the relationship between models and their originals ("systems") including the distinction of descriptive and prescriptive models, the view on various kinds of instantiation and the way how metamodels are defined. Some of these questions are debated in detail and alternative positions are presented.Keywords Modelling · Model and original · Prescriptive/descriptive model · Transformation · Projection · Type model · Token model · Ontological/linguistic instantiation · Metamodel

No abstract

No abstract

Since 1999, Rational's Unified Process (RUP) is being offered as a guideline for software projects using the Unified Modeling Language (UML). RUP has been advertised to be iterative, and incremental , use casedriven and architecture-centric. These claims are discussed while RUP core concepts like phase, iteration, discipline (formerly: workflow ) and milestone are reviewed in more detail. It turns out that the RUP constitutes a considerable step towards a broad dissemination of software process modelling ideas but some of the RUP definitions and structures lack clear structure and are too complex and overloaded for practical use.Among others, I see the following particular problems: (1) phases do still dominate the process and iteration structure, (2) the term "software architecture" is not clearly defined and its role is still underestimated, (3) RUP "disciplines" are a partly redundant concept complicating the process more than supporting it, (4) powerful and transparent structuring principles like recursion and orthogonality do not get the attention they deserve. As an alternative, our model for E volutionary, Object-oriented S oftware development (EOS) is contrasted with the RUP.

No abstract