Search

A computer program based method is described for creating models using gestures. On an input device, such as an electronic whiteboard, a user draws a gesture which is recognized by a computer program and interpreted relative to a predetermined meta-model. Based on the interpretation, an algorithm is assigned to the gesture drawn by the user. The executed algorithm may, for example, consist in creating a new model element, modifying an existing model element, or deleting an existing model element.

Distributed collaboration is becoming increasingly impor tant also in software development. Combined with an in creasing interest in experimental and agile approaches to software development, this poses challenges to tool sup port for software development. Specifically, tool support is needed for flexible, distributed collaboration. We intro duce the Distributed Knight tool that provides flexible and lightweight support for distributed collaboration in object oriented modelling. The Distributed Knight implementa tion builds crucially on the type-based publish/subscribe distributed communication paradigm, which provides an effective and natural abstraction for developing distributed collaboration tools.

Distributed applications are hard to understand, build, and evolve. The need for decoupling, flexibility, and heterogeneity in distributed collaboration tools present particular problems; for such applications, having the right abstractions and primitives for distributed communication becomes even more important. We present Distributed Knight, an extension to the Knight tool, for distributed, collaborative, and gesture-based object-oriented modelling. Distributed Knight was built using the type-based publish/subscribe paradigm. Based on this case, we argue that type-based publish/subscribe provides a natural and effective abstraction for developing distributed collaboration tools.

This paper presents linguistic primitives for publish/subscribe programming using events and objects. We integrate our primitives into a strongly typed object-oriented language through four mechanisms: (1) serialization, (2) multiple sub typing, (3) closures, and (4) deferred code evaluation. We illustrate our primitives through Java, showing how we have overcome its respective lacks. A pre-compiler transforms statements based on our publish/subscribe primitives into calls to specifically generated typed adapters, which resemble the typed stubs and skeletons generated by the rmic pre-compiler for remote method invocations in Java.

The emergence of the Unified Modeling Language (UML) has provided software developers with an effective and efficient shared language. However, UML is often too restrictive in initial, informal, and creative modelling, and it is in some cases not restrictive enough, e.g., for code generation. Based on user studies, we propose that tool and meta-level support for several levels of restriction in diagrams and models is needed. We furthermore present a tool, Knight, which supports several levels of restriction as well as ways of transferring models from one level of restriction to another. This approach potentially increases the usability of the UML, and thus ultimately leads to greater quality and adoption of UML models.

This paper reports from experience obtained through development of a prototype of a global customer service system in a project involving a large shipping company and a university research group. The research group had no previous knowledge of the complex business of shipping and had never worked together as a team, but developed a prototype that more than fulfilled the expectations of the shipping company. The prototype should:- complete the first major phase within 10 weeks, - be highly vertical illustrating future work practice, - continuously live up to new requirements from prototyping sessions with users, - evolve over a long period of time to contain more functionality - allow for 6-7 developers working intensively in parallel. Explicit focus on the software architecture and letting the architecture evolve with the prototype played a major role in resolving these conflicting constraints. Specifically allowing explicit restructuring phases when the architecture became problematic showed to be crucial.

We present three operators for the expression of event-based large-scale distributed programming. In short, these operators support a publish/subscribe-based interaction scheme, leading to a distributed object programming model merging the benefits of objects and events. For the integration of our operators into a strongly typed object-oriented language we identify a set of four concepts, roughly (1) serialization, (2) multiple subtyping, (3) multi-methods, and (4) closures, which provided by a language, enable its smooth integration with our model of publish/subscribe interaction. These concepts are illustrated through Java, which we augment by anonymous methods, a simple form of closures, and in which we use double dispatch to make up for the lack of multi-methods. A precompiler transforms code related to our publish/subscribe operators into calls to specifically generated typed adapters, which are similar to typed stubs/skeletons for remote method invocations.