Friday, August 11, 2017

Women in technology

The fact is that women are less able to code than men. This is from a survey of 36,000 individuals in India.


That means there are 4.45 men for every woman capable of writing correct code. Factor in sex preferences and there are probably 50 men capable of and interested in writing correct code for every woman.

12 comments:

Shimshon said...

I worked at eToys (in Los Angeles) for about a year from 1999 to 2000, during their peak employment. They had a pretty sizable development team for a company with around 1000 employees. I cannot remember a single female among us. It was a pretty talented group too. That was the largest post-college employer I ever worked at by far. I am grateful for having avoided the corporate EOE dogma as it metastasized into Stage 4 SJWism.

Tatooine Sharpshooters' Club said...

Funny, there was some choad on Twitter arguing that women's code "was judged as good or better than men's" when the gender of the coder was unknown. Shocker, he had no links, but this is not the first time I've heard this, and I would like to see this so-called study, just to uncover the scam.

Of course, the "equalists" always throw in the demonstrable untrue "or better", because if there were true blank-slate equality among all the diversatrons, then their entire project is for nought.

dc.sunsets said...

Not a programmer but have a question.

I'm told that coding is increasingly kind of scut-work, often handed off to code monkeys while the real work is at more of the architecture level.

Any truth to that? If so, then a relevant question is, who (men or women, WASPs or Wops, whites or blacks, etc., demonstrates the higher frequency of aptitude to think at this higher conceptual level?

My suspicion is that it's mostly white men who demonstrate the required level of intellect & drive.

Nathaniel said...

@dc.sunsets

They say that, but they've said that for 20 years. There are businesses that attempt to make a go of that structure - having a competent, US-based coder having to spend his day on Skype with the India or Philippines office, explaining how to code to people making $10/hr. It has led to an epidemic of really bad software.

People who can code - here - still make a lot of money. Architects are still people who can sort of code, but aren't the greatest at it, but can see the big picture and understand how the system components work. Managers are often those who are just mediocre at both, so, intellectually can understand it, but can't do it.

There's still this idea that you can have really poor quality code, and somehow, the system will save you. Moore's Law is at its end; there's not a great performance increase as fast as there was, since you can't make your chip any faster by adding more pipeline steps and cranking up the clock. Auto-parallelizing is very domain dependent, and you can't solve an inherently serial task with more cores. You can get better compilers, but, with more and more coding being done in high-level scripting languages, your compiler is not going to save you.

So, although I agree that coding is looked upon as scut-work, it's because there's still this fantasy that it doesn't require holding three concepts in your head at once - the problem you're solving, the library you're interfacing with, and how the machine itself actually works - which is a level of abstraction that the old "brain based learning" education movement of the 1990s would tell you is a level of abstraction only 7% of the population can pull off.

If coding were a solved problem, SQL injection would not still be a security concern 17 years after the allegedly-widespread adoption of things like DB.pm.

Schubox said...

I always go and read the references and this one is particularly pertinent to my professional life. Can you please provide a link to the reference? Thank you.

Tatooine Sharpshooters' Club said...

"Coding is just scut-work" sounds very much like C-suite talk to justify gutting the technical side to shareholders and pretending they're "adding value" by destroying the product.

Anonymous said...

Here's the whole thing: http://www.aspiringminds.com/sites/default/files/National%20Programming%20Skills%20Report%20-%20Engineers%202017%20-%20Report%20Brief.pdf

John D Alden said...

Funny, there was some choad on Twitter arguing that women's code "was judged as good or better than men's" when the gender of the coder was unknown. Shocker, he had no links, but this is not the first time I've heard this, and I would like to see this so-called study, just to uncover the scam.

There was this study a while back. That's probably what he's talking about.

I'm told that coding is increasingly kind of scut-work, often handed off to code monkeys while the real work is at more of the architecture level.

Any truth to that?


To echo what others have said, no, not if you're working on anything meaningful.

I used to be a scientist. Probably 95% of my job could've been done by any idiot with a few days worth of training and some reference sheets, at least at a particular site. They would've needed to be trained for each new site. However, the 5% left over was the really important part, the anomalies and edge cases that I was actually looking for. Since there was no guarantee what those odd samples would be, it required broad knowledge and experience to interpret them.

Well, programming is the same way. Maybe a lot of the code people write is essentially "insert round peg A into round hole A" type stuff, but when things go wrong or get weird, you need to know what you're doing to figure out what to do next. And that's just for writing code the way someone else told you to. If you're working from nothing but "solve problem _____" you really need to have a clue.

MegKelley said...

It's interesting to see these numbers because I was just thinking this about women coders this morning. Through school and the workforce, I have generally noticed that women are not as competent coders as their male counterparts. Obviously there are exceptions, but I've noticed that the women who are competent coders happen to be the rare woman who can sweep the floor with most anyone academically on any subject - not just outcode them. After a great deal of reflection I could only name two women I work with who are competent coders while with very little reflection I can name at least half a dozen men who are competent coders. I consider myself a poor coder but my one male coworker (poor guy) is anxious for me to return from maternity leave because the other members of our group are not up to snuff doing even the basic SQL our group requires. I'm really a bit frustrated with this because I've been giving one on one coaching for literally years now - but I can't deny his complaint.

Unknown said...

I'm 38 working in the software business for about 10 years, current a lead developer. Here's my take on it from my experience.

Women are very very rarely the very worst coders, but then they're never the best either.

The single biggest issue I personally find in dealing with women in software development is that I cannot give them negative feedback about anything in a clear and direct way without it causing problems.

Don't get me wrong, you get the occasional man with the same problem but this usually sorts itself out because HR/upper management are far less willing to step in and white knight for them.

Revelation Means Hope said...

Ha, what this study is really showing is that you should never hire India trained and/or India born coders.

Period.

A measly 6-7% chance that someone who is trained as a coder can actually code is a pretty poor percentage.

Harambe said...

Anyone who says writing code is easy or thinks of it as grunt work has never created any (successful) real-world software.

Post a Comment

NO ANONYMOUS COMMENTS.