Naresh Bhatia, Sapient Corp. (USA)
Mike Heffernan, OpusEdge (Canada)
Andrew Forward, U.of.Ottawa(Canada)
Ethan Hadar, CA labs and Irit Hadar, U.of.Haifa (Israel)
Stephen Nelson, Victoria U. of Wellington, (New Zealand)

Naresh Bhatia: AndroMDA – A Framework for Model Driven Development

Naresh Bhatia is a Senior Technical Architect at Sapient Corporation, a consulting company that helps clients innovate their business. He has architected enterprise applications for such diverse organizations as financial services firms, energy trading companies, universities and governmental agencies. Prior to this Naresh had founded a startup where he created one of the first UML tools. He is a committer on AndroMDA.

Abstract

The traditional approach to developing applications is very laborious and inefficient. We gather require-ments, create models, translate these models to code, and finally test and tune the “crude” application until it is refined. That's a lot of work! A large portion of this effort is repetitive, mundane and boring. In this session, Naresh will share his findings about significantly reducing the level of effort to deliver enterprise applications using AndroMDA, an open source framework for Model Driven Development and code generation.

AndroMDA has been used for everything from simple CRUD applications to complex enterprise applica-tions. It comes with an array of ready-made cartridges for common architectures like Spring, EJB, .NET, Hibernate, Struts and more. If you have a custom architecture, you can write your own cartridge to generate code that fits your needs.

The benefits of Model Driven Development include:
• Reduction in development time due to automated code generation
• Easier synchronization of application layers by using a single representation of business concepts
• Improved code quality due to encapsulation of best practices within code generation templates
• Support for system evolution as technologies come and go

In this session, we will use demos and code examples to illustrate how MDD can be used to simplify development and increase the speed of delivery.

Mike Heffernan: Associative Data Framework in C#

Mike Heffernan is CTO of OpusEdge Inc., an ISV in Ottawa, Canada. Formerly Director of Applied Technology Leadership at Nortel, in the 20+ years of his career, he has worked in such diverse problem spaces as code generation, development tools, network and browser based applications, customer relationship management, supply chain management, network management and telecommunications, retail point of sale and full text search technologies.

Abstract

Mike will describe a framework which enables rapid development and evolution of multi-user n-tier systems. In the heart of this framework is a model assembly which defines all application objects, associations and overall model policy.

This approach differs from Jiri and Martin's approach in that it is using meta-data attributes and reflection to dynamically bind an application model into the framework at run time. Application objects, associations and organizational classes are declared in a separate assembly, and multiple models can be loaded into the framework simultaneously. The approach reflects Mike's experience that explicit awareness of meta-information regarding the application's abstract model and associations at the code level is a key component in advancing software practices in the development of complex systems. The framework is implemented in C# on the dotNet 3.0 platform using a desktop WinForms client using full-duplex messaging via Windows Communications Foundation to a back end application tier that persists the data in SQL Server 2005. The model driven approach pervades every layer of the framework.

Andrew Forward: Actual Use of Modelling by Software Practitioners (a survey)

Andrew Forward is PhD Candidate at the University of Ottawa, Canada

Abstract

Despite both theoretical and demonstrated benefits of modeling with languages such as UML, many software engineers either continue using a code-centric approach, or use modeling but only half-heartedly. My research investigates the reasons for this; it also explores approaches that may help to bridge the gap between code-centric and model-centric thinking. This involves empirical studies into how modeling is perceived, practiced, or not practiced, and then an experiment with bridging the gap between modeling and programming in various contexts such as web applications, business process development, and thick-client applications.

I would like to hear the opinion of other participants on this topic. I will present my first study -- an Internet survey of software professionals, as well as my preliminary work on how to bridge the gap between modeling and programming in both Java and BPEL (Business Process Execution Language).

Ethan Hadar and Irit Hadar: The survival of the fittest -- the UML required evolution in weaving applications.

Ethan Hadar works for CA Labs, CA Inc. Yokneam, Israel, Irit Hadar is at the Department of Management Information Systems, University of Haifa, Israel.

Abstract

Click here to get the expanded abstract in the .doc format.

Stephen Nelson: First-class Relationships in Object Oriented Languages

Stephen Nelson is PhD Candidate at the Victoria University of Wellington, New Zealand.

Abstract

Stephen is preparing a thesis with the title First-class Relationships in Object Oriented Languages which is also one of the key issues we plan to discuss at the workshop. His position is described in the abstract submitted to another conference.

Return to the workshop HOME PAGE