Some disjointed thoughts I've had recently that are bound together by the notion of workshop-philosophy; the phrase is formed on the analogy of 'workshop-art', i.e., the art in the workshop itself, including sketches, half-formed ideas, first attempts, etc., as well as finished works.
Jorge Gracia, in A Theory of Textuality, suggests a useful terminology for talking about commentary-type 'interpretations'. We have an interpretandum, the text to be interpreted, and the interpretans, text which is added to the interpretandum by the interpreter, and the union of the two is what we often call 'the interpretation'. ('Text' here would include spoken as well as written use of signs.) So, for instance, 'Averroes's interpretation of Aristotle' is Aristotle (as Averroes would have had him) plus Averroes's text added to Aristotle's text. Seen in this light, the crafting of a good interpretans is a major part of what philosophers do: even if you are simply discussing an argument made by a contemporary, what you are doing is creating an interpretans to go with it.
An interpretans must be crafted; it is like a work of art. Part of its relation to the interpretandum is merely logical -- one aims at consistency between interpretans and interpretandum; but no one settles for this, either. What seems to be added above this is a set of aesthetic criteria -- a sort of theory of genius and good taste (to use the early modern terms), or at least rules of thumb and rought approximations tending toward such a theory. So, for instance, we like interpretantes that add new harmony, but prefer this new harmony to grow organically (as it were) out of the interpretandum, just as we find marble sculptures striking and impressive where they seem to grow naturally and easily out of the marble itself. Integrity or completeness, proportion or consonance, clarity or splendor.
If evaluation of an interpretans is partly a matter of critique or good taste, then the way to go about becoming good at it is the same as with every other kind of good taste:
(1) Develop a broad base of relevant experience, so you can make informed comparisons;
(2) Build the relevant skills of discernment, i.e., the acquired ability to identify important moves, novel twists, and the like;
(3) Work toward having good sense, understood as the self-critical fairmindedness that allows us to be objective and unbiased.
And then, perhaps, the good crafting itself, described by the theory of genius, is a matter of 'active taste', just as evaluation is 'passive genius' (to use Beattie's terms). Or else (as in Kant), it is giving rule to art, the idea of which the evaluator needs in order to evaluate the work in an appropriate way. Or else (as in Schiller), play, i.e., the free play of mind and imagination in such a way as to unify reason and sense, freedom and necessity.
Novalis suggests the idea of a poetry that is in the workshop: "Stories can be thought up that lack coherence, but have associations, like dreams; poems that are merely melodious, full of lovely words, yet without any sense and coherence, only single stanzas understandable, like fragments from diverse things. This true poetry can have at most an overall allegorical sense, an indirect effect, like music has. This is why nature is so purely poetic, like a magician's workshop, a laboratory, a nursery, a carpenter's storeroom." Similarly we can think of a kind of philosophical work that is still in this dream-stage, "fragments from diverse things," suggestive like music of something greater, full of endless potential like nature in its poetry. Here is an argument, there an idea, here a redaction, there a question, here a possible line of inquiry, all piled together in the storeroom, in the back of the mechanic's shop, until they might be needed. And this would not be any less philosophical than the fully developed work; it is, in fact, the beginnings of it. And if philosophy is done properly, this magician's workshop, this complicated laboratory, is there; we all have it, but simply ignore it except in rare cases where some of the results of the workshop become famous. Then, just as even Dickinson's less successful attempts or Coleridge's marginalia become unusually interesting when Dickinson or Coleridge are famous, so too does a notebook by Hume, or a scrap of conversation by Wittgenstein, or even the catalog of someone's library, become interesting when they do something that becomes famous. But the workshop was there regardless; if all of Dickinson's truly great poetry were lost, we might still have bits of her poetic workshop. It would be less obviously interesting, as a sketch by Leonardo would be less obviously interesting were it the only thing by him that it survived, but it would still be the work of genius, the workshop-work of genius. Master craftsmen don't create ex nihilo; they select their materials, their tools, their purposes, try things out and consult with others, and this is as much art as the final finishing. And so too with philosophy.