Scientific writing is writing about science.
Scientific writing in English started in the 14th century.
The Royal Society established good practice for scientific writing. Founder member Thomas Sprat wrote on the importance of plain and accurate description rather than rhetorical flourishes in his History of the Royal Society of London. Robert Boyle emphasized the importance of not boring the reader with a dull, flat style.
Because most scientific journals accept manuscripts only in English, an entire industry has developed to help non-native English speaking authors improve their text before submission. It is just now becoming an accepted practice to utilize the benefits of these services. This is making it easier for scientists to focus on their research and still get published in top journals.
Besides the customary readability tests, software tools relying on Natural Language Processing to analyze text help writer scientists evaluate the quality of their manuscripts prior to submission to a journal. SWAN, a Java app written by researchers from the University of Eastern Finland is such a tool. 
Writing style guides
Different fields have different conventions for writing style, and individual journals within a field usually have their own style guides.
Some style guides for scientific writing recommend against use of the passive voice, while some encourage it. Some journals prefer using "we" rather than "I" as personal pronoun. Note that "we" sometimes includes the reader, for example in mathematical deductions. Publication of research results is the global measure used by all disciplines to gauge a scientist s level of success.
In the mathematical sciences, it is customary to report in the present tense.
In the chemical sciences, drawing chemistry is as fundamental as writing chemistry. The point is clearly made by 1981 Nobel Prize in Chemistry Roald Hoffmann.