thinking briefly on annie finch's idea of "multiformalism" this morning, together with my idea of "transposition"
in systems (and a poem is a little system, or a piece of one (if it is a serial poem, say), or a version or release or iteration of one), one thing you have to consider is the formalism of your approach to the system; if it is a little poem, will it bludgeon it to death and cost a lot of irrelevant effort and overhead to use a very formal process, or will something important occur as the output?
well, in this case, learning how to make a system properly is a good -- if you're writing the poem to learn to write poems, then a certain amount of formalism may be a benefit -- not immediately but to systems down the line -- but in a piece of something else, you want just the right amount of formalism -- never done a use case, and got some hairy reqs? do the cases, and see where the chips fall -- or shlep the cases through the process if they've been getting abandoned as early lifecycle decoration during data modelling --
so, too, ... lost track
opportunities for disruption and variation seem -- not just offered by form by also by semantics, by grammar, by subject matter (being multidimensional)