There is something a bit "off" in the question, so it's not clear what you are looking for. It is conceptually wrong to historically "change" Norwegian into German, but you can meaningfully compare Norwegian and German (because they both descended from the same earlier language). You can even give correspondence rules between Norwegian and German (although language-to-language comparison tends to be unreliable if the common language was rather far back in the past). If you just want "comparative sets", I recommend Malcolm Guthrie's Comparative Bantu, which assembles masses of data and has a keying system that allows you to make language-to-language comparisons. (That works by indexing the data to something that is in fact homologous to a historical reconstruction, though he says it is not one). The first volume is full of methodological hints.
One way in which you can meaningfully write rules changing one language into another is if the first language actually is the historical ancestor of the other, for instance Middle English and Modern English, or Latin and French. In the latter kind of case, in practice we do not rely just on the facts of French, we look at all of the Romance languages, and that is really how we get the chronology right. The reason for that is, more or less, that the language that Romance languages descended from is not (classical) Latin itself, but a later spoken version that differs from Classical Latin in a number of ways. As an exercise, an instructor might give you words of Proto-Salishan and words of Lushootseed, and ask you to write out the rules (sound changes) that get you from the former to the latter – except, we have no records of Proto-Salishan, and are not really sure about all of the details of the language.
Generally, we use the comparative method to figure out what the earlier language is and then figure out the steps involved in getting to specific languages. This requires looking at a number of languages to create a third previously unknown language; so you could look at Norwegian, Swedish, Dutch, German and English and try to come up with something about earlier languages.
It looks like what you're looking for is some version of the basic methodology of solving phonology problems, going from underlying to surface forms. There are a number of textbooks out there that do this, though generally they take it to be self-evident how you come up with rules and order them. You might try Introducing phonology which emphasizes methods, although it requires you to figure out what the initial state is (and justify that belief), so it's really more like doing comparative work.