@postmodern
That might be a useful #metric. How often, if at all, are these #constant #regexp lines likely to be edited? How many are #autogenerated? Will they ever need to be #regenerated based on new knowledge or #constraints? Rather than thinking of #partitioning them in a #functional or #elegant way, how about in terms of sheer #maintainability? What's easiest for #editing, #documenting, and #maintsining?