I found this research report disappointing and superfluous. It rather reminds me of a midterm graduate or internal working paper. Having been attracted by the interesting title, I found myself wondering why this collective of seven authors spectacularly failed to deliver almost any (I am sorry to say) tangible result.
My best guess is that the authors tried to cover too much ground and, in the end, could not deliver at all what they promised.
For instance, they touch on too many workflow-relevant domains, ranging from model-driven architecture (MDA) to semantic data extraction and mapping using the resource description framework (RDF). Despite the broad range of topics, the subsequent analysis and exposition stays on a trivial level, such as when they introduce Extensible Markup Language (XML) syntax. A meta-model of their purported architecture (using unified modeling language (UML)) clearly shows the immaturity of their work (design-time activities are not linked to runtime tasks; a task list can only serve a single employee, even though “roles” are a key element of the title and abstract; and a corresponding entity is altogether missing). All of this is exacerbated by sloppy, sometimes confusing English and a lack of rigorous definitions.
In the end, the authors claim that they have used their framework to implement not a workflow engine, but a simulator. But instead of disclosing the internal workings of their simulation engine, they only show a barely legible screen shot of a design-time process model and another diagram (of equally low quality) with some numbers derived by ex post analysis of a simulated process.
Personally, I would have never accepted such a paper in the first place. My best recommendation is to ignore the work, unconditionally and completely.