Thoughts from our workshop AMINO at MODELS 2013
There is certainly more to models than just code generation and we were happy to see several participants at MODELS 2013 echoing this. We were also glad to see that presentations at AMINO workshop fostered the feeling that models indeed, could be used in a variety of ways towards modeling an organization.
In an earlier post, we indicated that in general, models have been used to capture the “what” and “how” of business applications. We are pursuing one important dimension that is left out, but which we think is the cornerstone of decision making in organizations, the “why” dimension.
The AMINO workshop keynote focused on this dimension stating that not only what, how, but why too, are all important aspects to consider in decision making.
Architecture descriptions or higher level specifications should reveal intentions/goals rather than just the facts about systems being modeled. Goals, capturing the why behind design decisions, could in fact, be treated as the focal point with multiple interrelated models at every level of organization. Modeling of goals could be simplified by understanding that goals are anchored in properties of domain elements.
Other researchers proposed to capture conflicts amongst goals using some form of argumentation between modelers and stakeholders where counter arguments challenge assumptions. Dependencies between different kinds of data, processes, and process stakeholders require special treatment, as changes in the enterprise context affect these and other myriad aspects of enterprise models. The domain assumptions should be monitored for their continued validity as goal decomposition may be affected by change in the environment within which an organization operates. The rationale modeled at higher levels of abstraction should eventually reflect in IT systems being generated/ modified/ analyzed. Discussions at AMINO concluded that representation of rationale is required for realizing business intent wherein multiplicity of models is a reality.
GEMOC representatives held a cameo at the end of AMINO workshop, where the need for heterogeneous model-driven engineering was accentuated. It would take the form of modeling language engineering capitalizing on relationships between multiple modeling languages. The focus on purposive modeling languages including domain-specific languages (DSLs) was evident also in DSLFin workshop where DSLs for expressing and analyzing financial contracts, for high performance and parallel computing, and for taming the complexity of messaging were described. The aim of Multi-paradigm modeling workshop seemed similar in spirit to that of AMINO – focusing on the need of concise, view-specific, heterogeneous models in different languages to fit together.
A number of interesting keynotes and panel at the main conference fueled discussion centered on abstractions and models used to capture those abstractions, be they models of enterprise or models of code itself. The need to educate students and programmers about capturing abstractions was brought out and various ways were discussed for teaching abstraction, whether as a process of removing irrelevant details and/or as a process of capturing commonalities and patterns across similar situations. Some tutorials brought out the fact that abstractions are dependent on viewpoints.
Various main conference presentations revolved around using modeling and metamodeling issues. On behalf of TCS, two presentations were made, one focused on issues in model versioning and the other on using suitable representation for making models of enterprise machine-processable and analyzable.
All in all, it was good to see a rising awareness about using models for purposes other than/in addition to code generation with many showing interest in our vision in general and our workshop in particular.
P.S. Tweets streamed during the AMINO workshop can be found on twitter using #seekAMINO.