From Lazy Modeling – Aspects and Examples by Jari Peltonen:
“We believe that being flexible means essentially the same as being able to adapt to the prevailing circumstances of the work. The current modelling tools are not flexible enough in this sense. A general assumption seems to be that the work is adapted to the tool, instead of doing it the other way around. All the decisions about used tools, notations, ways of working, etc. have to be done beforehand, in the beginning of the project. This is unrealistic and impractical. Hence, we propose a “lazy modelling” approach in this paper, that is, support for just-in-time decision-making and late binding of tools, practices, notations, etc. We also consider what kind of tool support is needed for the approach to work in practice, mainly through our existing prototype implementation.”
I deeply agree to the analysis of the situation in the paper. The subsequent approach of the authors to address these issues is strongly focused on how an appropriate tool may look like. Certainly, providing appropriate tools is a reasonable objective, although I think on their way the authors are omitting an important step. I could not say crisply what kind of result I would have expected, but an investigation, focused on how the underlying logical structures develop during project progress and how tools suite to their different kinds/ complexities, would have sound very intriguing to me. Perhaps roughly sth like: “in the beginning we found that requirements consist mainly of grouped lists, growing to tree-like structures and finally become Boolean matrices. Up to trees you’re doing well with word and as soon as matrices are involved excel like tools are a reasonable choice and if they contain not too many relationships an ER-Modelling tool is somehow helpful …”.
So far, just some ideas.
also notice Vincent Hanniet’s example here