Replacing mankind with humankind, chairman with chairperson, and manpower with workforce (or with such equivalents as "human resources," "labour force," and "personnel") is now routine: even if some writers use gender-specific terms, copy editors replace them with suitable gender-neutral substitutes.
However, what if the text requires gender-specific terms? In medical writing, it often needs to be specified whether the patients or subjects were men or women. In ecology, wildlife biology, and experimental zoology – domains that require that the sex of the animal in question be specified – the words female and male are routinely used, and the usage poses no problems. It is only when writers need to choose between female and woman that the choice becomes tricky (choosing between male and man is seldom a problem).
Fortunately, the advice offered by most style guides is straightforward: use female only when the biological distinction is relevant or needs to be preserved, as in female secondary sexual characteristics, female preferences that govern the choice of a mate, or calories required by nursing females. Deborah Tannen, a sociolinguist who has written a great deal about gender differences related to the use of language, says that she avoids "using “female” because it feels more like describing an animal than a person." In other situations, assuming that a distinction is necessary, use women (as in women subscribers, policewomen, women teachers, and so on).
What is not straightforward is the choice between making the distinction and not making it: it appears that readers, especially if they happen to be women, are more likely to be annoyed if the distinction is made without good reason than by the word – woman or female – chosen by the writer in making that distinction. Even harder to explain is the illogical but common pairing of woman and male as the adjectives of choice when specifying the gender: Maddie York quotes the following recent headline from the Guardian "Women bosses earn 35% less than male colleagues." This pairing is encountered in academic writing too: Google Scholar, for example, when searched using the string [women*male], retrieved 3.6 million results but only 3.1 million when the string was changed to [women*men].
You might also be interested in knowing Should I use the word "sex" or "gender" in my scientific research paper?