Roger, OK, this should be more specific. For manuscipts for publication, first you have to do the research including reading up to several dozen published papers. Then, it is a good idea to make at least rough versions of all the figures, so you know what you are writing to. Then, write. There is a particular format: first a short abstract (although you might write that last). Then, an introduction, which is important. You set the geographic and geologic background. You discuss previous interpretations and models from the literature, but not in "book report" form. Instead, you might talk about one aspect of the interpretation. For example, in the manuscript I am working on, several published papers say that the North Anatolia fault did not form as a discrete single main fault until about 200,000 years ago (Imren et al, 2001, Demirbag et al., 2002, Le Pichon et al 2001, Sengor, 2005). Before then, it was series of normal faults bounding basins linked by transfer faults.

In contrast, Armijo et al 2002 has a steady-state extensional model. Seeber et al 2006 (includes HW and I) also do not see any significant changes in the last million years or so, but their interpretation is very different (although I have to re-read this).

So, you set the controversy, but try not to be long winded about it. You also normally do not say what your interpretation/model is yet. Inexperienced science writers (like students) tend to mix text on their interpretations with text on what others think that is wrong...this gets confuding really quickly.
Many who have only been working in one area (say; southern California) write in a style that is "of local interest" and these are often rejected by the more prestigious journals. You want someone from New Zealand to be interested in your manuscript on, say, the San Jacinto fault.

Then, you have a section on what you did to address the controversy. You talk about data and methods.

then, you talk about results: "this is what we interpret". But, for offshore faults and seismic reflection data, it is NOT good to describe each seismic line and each fault: becomes unreadable. Instead, you make high quality figures that show what you are talking about; then you need less words.

Then, you have a Discussion section. This is where you explain why your interpretation and model is correct, why certain others cannot be correct, and also talk about what you have done means: why it is important (address questions raised in the introduction).

You then usually have a Conclusions section: your main important conclusions.

Then, list of references.

Then, you have to have it in the format of the Journal you are submitting to.

Then you submit. The Editor assigns the manuscript to an assistant editor (for certain journals). The editor then finds reviewers: there are 2. Eventually (a couple of months if you are lucky), the reviewers retrun their reviews to Editor or assistant editor; if there is an assistant editor, he/she makes a recommendation and may do an additional review. Then, the editor decides whether to reject, or ask for major revisions and resubmittal, or accept pending minor revisions.

Chris