Modeling for real variety implies that the modeler cannot stop at choosing and partitioning a single situated object (analysis). S/he has to integrate (synthesis) a host of such choices. The results of analysis must already be equipped for synthesis, with synthesis a test of proper analysis. That is, analysis and synthesis are not antithetical. Instead, they are … complementary. The modeler has to ‘constantly’ iterate her/his mode.

in: Metapattern for complementarity modeling

 

On purpose I’ve tried to suggest a ridiculously generalized conceptual model. […] I am sure you want, and therefore should, move concepts and/or relationships around. Please do! At this stage, i.e., of conceptual modelling, I find the challenge is not at all technological (if ever). So, it doesn’t look much like work, not what most people consider real work, that is, because you are not seen to be busy building et cetera. However, what (meta)conceptual compactness remains valid after prolonged scrutiny — and it does take time for imagination to take effect; the back of our mind, whatever that is, is where imagination is most productive — provides exponential benefits. So-called constitutional relationships where some type is retyped into another type for — yet — another cycle of instances, and so on, help do the trick of recursion (which limits the model while expanding the possibilities it offers).

in: note 71.33

 

Without a tool supplying the false illusion of control and preferably even seamless relevance for programming, a conceptual modeler is forced to focus on … conceptual structure. Whatever tool should not distract or, even far worse, steer her or him.

in: note 71.40