There is always more to read, but I currently see it as a step-wise process of non-derivative hypothesis generation, each from a different source.
1. from data. (thesis)
2. from a counter-example to 1. (anti-thesis)
3. from substructural-negations of 1 and 2. (synthesis)
I put non-derivative in italics because I think it's very important. Many people make the mistake of thinking synthesis is a compromise between the other two hypotheses when it should be something novel, innovative, and non-derivative which subsumes their substructural negations. To a lesser degree some people think that anti-thesis is a mere substructural negation of the thesis, "exception-barring" as Lakatos calls it, but it should be just as exceptional as the first thesis.