The reason female participation isn't larger is because of the education / training part of the pipeline. Having a welcoming workplace doesn't suddenly spawn people with skills wanting to work there.
Rather than being about a male-dominated "lone coder", Microsoft's guarantee of a private office means two things:
1) that a private office is a benefit to their target employees, and
2) that most companies at the time did not have a private office
Otherwise, they wouldn't advertise it as much. In more recent times, I don't think that Facebook's push to open offices is as employee centric as Microsoft's push to private offices. Instead, I think Facebook (and Apple) are trying to save money, since an 8' x 10' office in San Fran costs an employer anywhere from $8,000 / yr to $30,000 / yr. So, it would have less to do with providing benefits to their employees and more about their own bottom line.
Based on that, I'll say that measuring actual work in cognitively focused fields is difficult if not impossible.
The entire premise of compensation based on or a requirement built around a given minimum or maximum hours of work is crazy when the actual work that happens is mostly cognitive and not literally constructive. A job of that type practically demands a salaried payout and advancement based on overall project success/contributions (which might be incredibly fuzzy to measure and difficult to prove or compare as facts).
> "We’re now in an era where computer scientists are no longer universally considered to be ‘80s-style loners, but the structure of workplaces is still often dominated by the manner of thinking that was created by them (see also: “move fast and break things”[1])."
So why isn't female participation coming back up? It was higher in the 1970s and 1980s than it is now.