The structure of the Language
Discussion on the purplsoc paper
The structure of the language
The PLAST will actually not be built on the name, context, problem, forces, solution, action format of classical pattern language. It will consist of something more fine grained, to parse and analyze systemic experience in order to solve what Jenny Quillien calls problems of organized complexity, arising from heterogeneous components interacting nonlinearly, giving rise to new behaviors. Jenny Quillien quotes Jane Jacobs and the need to identify interacting variables and to separate these variables in ‘smaller segments’ because they “do not exhibit one problem which if understood explains all. They can be analyzed into many such problems or segments which are also related with one another.” And, “when the segments are separated out the behaviors of a variable when in the presence of other variables can be discerned.”
A visual grammatized language provides good possibilities for discerning fine grain segments of behavior. In addition it provides good underpinnings for the writing of problem solutions type of patterns and the visual rendering of the dynamics of existing problem solution types of pattern languages. So it would enable the construction of both a priori and ad hoc contextualized problem/solution couples. In this respect, the PLAST is also a language for writing pattern languages.
 | Kurt Laitner 7:07 PM Nov 5 |
|
do you have an outline of what the metamodel is? from your napkin sketch you have defined some categories of Directions of Flow, Impulse, Process, Accumulation (pace and type). I am not entirely happy with the set or the factoring (for example, I think your Flow Directions is actually a combination of the generator of flow (an interesting bit that push pull crossed with directionality), cardinality (one to one, one to many etc), and directionality - mostly concepts from comp sci modelling (UML) and graph theory where referencing those domains may allow you to shortcut discovery of the basic set of operators (from UML, types of relations like association, aggregation, composition; from graph theory concepts such as cyclical versus acyclical (partly represented in your nature of impulse section), directed vs non-directed, weighted edges, multigraphs vs simple vs multidigraphs) - Not that I think you should necessarily adopt the language of these specialties but you should be aware of them and use the classification/ factorings that are available to you to pick and choose from
I will be able to help to some extent here, though graph theory is still a hobby, not a skill for me