Schneier on Security

Clive Robinson • May 26, 2017 12:43 PM

@ Alice Hutchings,

It’s nice of you to drop in here during what must be a hectic time at the moment.

The issues of gender imbalance in technology are as old as technology it’s self but sadly appear to have defied all attempts to redress it much so far.

I’ve been actively involved in a number of initiatives from the late 1970’s through to recent times in both the electronics and software industries, to improve the prospects for women. The one thing I can say is that women most certainly do not lack the academic or more practical skill sets, in fact they are often superior in the pre and graduate stages.

You are probably aware of Freeman Dyson’s “Bird or Frog” view of the way people think are categorized. For some reason when it comes to working in the software industry frog thinkers tend to be over represented when people talk of “Star programmers”, however when it comes to management above the team leader level bird thinkers start to predominate. Which is one reason given as to why few star programmers ever move up.

There is however a problematical social expectation that women should be “multitasking” though as you probably know many of the studies into multi / mono tasking/thinking do not show a general advantage to either mode.

Thus there is a societal in built pre disposition that the software industry needs monotasking deeply focused frog thinkers, not multitasking broadly focused bird thinkers, therefore the software industry is often perceived as,”No job for a women”.

This is actually far from true. Frog thinking tends not to scale for a couple of reasons, firstly whilst it has depth it lacks breadth and secondly it tends to preclude the ability of effective communication which is vital for projects beyond a small scale. When you look at extrodinaraly large projects in the software industry many tend to fail expensively. The reason usually given is “lack of communication / cooperation” between the disparate parts of the project. That is the projects are not effectively Project Managed.

The easy conclusion is “frogs do not become princes” no matter how hard you promote them, where as “birds take to the sky and fly high”. But it does not resolve the actual issues.

As is now being found at non Nation State level ICT projects, monolithic code constructs do not scale beyond a certain point even with multiple load balanced instances. Thus the drive towards microservices via Docker and Kubernetes. But in startups and smaller organisations the switch from monolithic to microservices is usually a disaster. There are two problem areas the technical and the human. Usually it’s the technical that gets blamed when in fact it’s the human side that fails, usually because of non multitasking and the lack of communication that arises.

I’m known for my preference for Hard Science post grads not CompSci for various reasons. Partly it’s because hard science post grads tend to be scientists/engineer oriented not artisan/copyist, but importantly they generally know, not just how to communicate, but also integrate disparate parts as a norm not an exception. Further and importantly from a number of peoples points of view women tend to be better represented in hard science.

But further when you look at people in CompSci at the starting positions in the software industry if they are frog thinkers they fit in, if bird they tend not. Thus you have the problem that there is a high attrition rate at the normal entry rung on the career ladder. But it gets worse, as frog thinkers go up the ladder they bring other frog thinkers in behind them not bird thinkers. Which because men predominate in the frog thinking at the entry rung tends to mean that frog thinking women pull in frog thinking men behind them due to the paucity of frog thinking women… Which means that the paucity of bird thinkers above the team leader management, who also have entry level experience is very very low anyway and vanishingly small when it comes to women.

Thus in my long but somewhat limited experience you need a different way for bird thinkers to get the entry level experience but enter in above the team leader management level. Then find a way to get bird thinkers not to leave at the entry level rung as they can then see a clear career progression opportunity.

The problem is that currently the software industry is set up around the idea of monolithic systems, even when compartmented the view is single instance orientated, not distributed component orientated. It both favours frog thinkers and worse locks the industry into an evolutionary cul de sac where the dead end is the laws of physics for a single processing instance.

As can be seen with CPU’s that are now multi-core the next evolutionary stage is parallel processing, which can only move forward via distributed systems. Whilst the design of single instances favours frog thinkers, multiple instances require a different thought process that favours bird thinkers who can multitask and communicate effectively. Thus as we move forward in the ICT industry, science / engineering not artisanal / copying will be the way forward which should favour bird thinkers more and more. Which as the frog thinkers in general will not be able to step up easily will actually create favourable opportunities not just to bird thinkers but women who tend to be better at the skill sets required at a younger age. If there is a sufficient skill shortage gender discrimination should diminish if not disappear due to supply and demand, and as older more staid / stale view points retire or fail to rise. But I must say that with over a third of a century industry experience so far the rate of change appears to have been less than the growth rate of the industry…

There is also the issues of the two world wars to consider. Both showed that outside of certain strenuous manual tasks women were just as capable as men and in some skills rather more so. However after each war women either left or were forced out of the work place, for what were mainly sociological reasons. It was not until the late 60’s through to the 80’s that “working wives” became more acceptable. Further in the US in the mid 90’s for a decade or so, in some ethnic groups women on average had higher pay rates than men, but caution is needed as the same ethnic group showed a much much wider educational achievement between the women and men, with the men entering low skill or working class employment, whilst the women tended to enter middle class better paid administrative and lower / middle management roles.

However there is now another way by which gender neutrality may occur, which is distance working. It’s not just the “nobody knows who you are, just what you do” aspect, it’s the ability to more easily encompass a much wider range of life styles. Importantly it actively favours those with good communications skills amongst many often disparate entities which favours the bird thinkers over the frog thinkers. However when we look at major FOSS projects we still see a male predominance. However we also see a higher predominance of monolithic projects which still favour frog thinkers, so this may well change as the need for parallel / distributed systems rises.

Finally, whilst there is sexism in the software industry, it’s also in many other work domains as well. It further tends to be more prominent in more polarised environments. In effect as a clique forms societal norms change, normally toward a common denominator and in short order a clique develops a “them and us” world view. Once developed separating the members of a clique does not immediately normalise the societal norms and if members meet again they will for some time revert to some extent back to the them and us viewpoint. Which makes undesirable traits linger.