When I came into the trade in the mid-1960s, in the corporate offices where I worked about two-fifths of the programmers were women -- in other words, a higher proportion than in the workforce of that time as a whole. In those days programmers were taken off the street and out of completely different occupations, since no one knew how to create a programmer and women were as good at it as men -- or maybe better. To my perception they tended to be the ones who got the work done whereas men often preferred to play games and show off. (But showing off and playing games can be useful learning experiences on a new terrain.)
Serious male domination of the field began when academic institutions tuned into the status, power and money which the programmers were acquiring and turned the craft into "computer science" and "software engineering." Academic science and engineering in the 1970s filtered strongly for maleness. The status and place of women in computer-related occupations (other than keypuncher or data-entry clerk) then declined into a Problem and an Issue for the same elites who had produced the situation.
During both periods, in any case, the main energies of development were directed by the needs of Capital -- women did not write a different kind of software from men, although in the early days, when programming was an incomprehensible art to most people, they were probably free to. Just so, today women are free to design all the web sites they want, at least if they have a few hundred dollars to set one up, but the design of "important" web sites and other software, where people receive a lot of money for the work, is probably going to be managed by the agents of Capital according to their perception of Capital's needs, as in the past. This has been true in every case of corporate web sites and software development that I know about. The question of gender doesn't enter into it, unless you say Capital is a gender -- maybe an interesting thought.
Gordon gcf at panix.com http://www.etaoin.com