In an issue from december 1991 of New Scientist, there is a curious article entitled "How to design an alien". It's definitely not the astrobiological dimension that struck me as relevant, but rather these quotes:

"[inventing new planets] requires building a new world ecology with a detailed evolutionary history so that the plot does not have any obvious contradictions (...) although every detail must be different, there are patterns of general problems, and common solutions to those problems, that would apply to life anywhere in the Universe"

Why do I blog this The "theoretically possible after all" dimension as well as the balance between particular challenges/universal solutions described in Cohen's article is very interesting, as an exemplification of the design process.