I think he misses the point of at least some of the criticism. It's not a matter of fairness, it's a question of whether it's a good idea for employers. You want to exclude all of the brilliant engineers who work on proprietary software? Go right ahead. But then don't turn around and complain about the "shortage of qualified engineers."
This part of a broader problem with the industry. Tech companies subscribe to stupid hiring fads (see Google's earlier reliance on brain teasers) and apply hyper-specific screening criteria (10+ years of .NET experience), then bitch and moan when they have winnowed their applicant pool to nothing and can't meet their recruitment goals.
None of my code is available to be open sourced, our github repo's are private (when sharing is required with outside parties) or self hosted internally. If a company looks at my github, you'll see a few contributions to some projects but very minor and few.
If you're making the decision to hire me based on my public github, you're making a mistake.
Indeed. I've been using github lately too, writing code for a charity even, but it's not code that can or should be open sourced.
Github-ism is a bit like credentialism, it lowers your risk but risk is where the interesting stuff happens.
There are two ways to hire, the lazy risk-minimizing way and the non-lazy investment minded way. If you put a maximum amount of effort into trying to reduce the risk of hiring you're going to cut your potential RoI off at the knees. You only hire people that are currently employed, because you want to rely on someone else's work in evaluating a potential employee. You only hire people with the right credentials who have relevant experience in exactly the technologies your company is using right now. You only hire people with very explicit demonstrations of their coding capability. Candidates that have such an easy time proving themselves are naturally going to be the easiest to hire. Not just by you, by everyone. Which means it's going to be more expensive to acquire them.
You should still acquire employees such as these, but if that's all you're doing you're throwing away a lot of potential. If you can manage to put together enough internal talent to be able to interview candidates even remotely well enough to make an independent and more or less objective check on their skills and capabilities then that gives you a potentially very significant competitive advantage.
Some of the best coders I've worked with came from diverse backgrounds and didn't have the "right" credentials or resumes. Several of the smartest and proficient programmers I've worked with have not had college degrees. Many others didn't have CS degrees.
Being able to find these folks and hire them because you can tell they have talent is a great opportunity to extract RoI from the hiring process, which is normally seen as just a cost-center.
I have to completely agree, my public profile only shows 1% of my ability. Its a double edge sword, either spend more time fluffing up your public profile or working on something that actually makes a difference. Now that doesn't say don't do random hacks in your free time, but everything you do in your free time isn't open source worthy.
I agree about the fairness. I personally don't even have a GitHub account, because I have little use for one. My code falls to at least 95% in one of two categories: code I wrote for other people for money which I am not allowed to share or code I wrote for myself for a very specific purpose where no fitting solution already exists and that has little to no value for anyone but me. Sharing the latter would mean investing time in cleaning up and documenting things although nobody apart from me is ever going to use it. I don't see that as a good use of my time just to gain bragging rights.
Companies have no obligation to be fair in hiring (accept for certain legally protected classes). They have an obligation to run the best hiring process they know how in order to acquire great people.
"code I wrote for myself for a very specific purpose where no fitting solution already exists and that has little to no value for anyone but me."
I always assumed public repositories are for these types of projects too. "Hey this code isn't clean or maybe finished but it solves this really specific problem maybe it could help someone". I understand being nervous about not putting your best code out there but I think there is also something to be said for releasing your problem solving attempts.
Also, in the first paragraph of the post "github or other code portfolio". The whole criticism was with respect to employers who only use github as a filter.
Who? The original article was written by someone who clearly identifies as "he" - Lee Carp is one of the two cofounders of BayesianWitch. He graduated from the University of Chicago...
That's bizarre, when I first looked at it, I skimmed it and at the end there was a picture of who I presumed was the author who was a male with red framed glasses which is now missing.
There is what looks like A-B testing code built into the page (but now configged off) - you can still see it in action in Google's cache of the blog posts or if you view source.
Option A: photoa.jpg Lee Carp is one of the two cofounders of BayesianWitch. He graduated from the University of Chicago with an AB in Sociology - specializing in understanding how technology shapes culture. He is fascinated by how marketing, technology and data intersect. He likes to swim and bake in his spare time.
http://www.bayesianwitch.com/media/img/photob.jpg
Option B: photob.jpg Shana Carp is the two cofounders of BayesianWitch. She graduated from the University of Chicago with an AB in Visual Art - specializing in understanding how technology shapes culture. She is fascinated by how marketing, technology and data intersect. She likes to swim and bake in her spare time.
http://www.bayesianwitch.com/media/img/photob.jpg
However, the domain is registered to Chris Stucchio, the HTTP metadata shows the S3 account name stucchio, the user making this HN post links to Chris Stucchio's weblog, and in that blog he isn't averse to spicing things up with an occasional picture of a scantily clad lady [1]
Whether bayesianwitch is a company with one, two, three or more employees we can only speculate!
What a load of asinine drivel. Posts like this never fail to upset me.
What's my excuse you ask?
I write highly proprietary software for algorithmic trading 10 hours a day and have zero time and inclination to work on an open source project in whatever little free time I have. I'd rather spend that time on hanging out with my buddies or girlfriend, hobbies, travel etc. Life is too short to waste solely on work and work related activities.
> You can come up with excuses all day about why you can’t show me your code. I’m going to hire the person who doesn’t make excuses and gets the job done.
Well good luck to you then, judging by the way you talk, it seems that you likely run a sweatshop and are screening for submissive people who are willing to work for free and bend over backwards. I wouldn't work for you even if my livelihood depended on it and I'm sure there are a lot of other self respecting developers out there who feel the same. Though we'd probably never meet because there is likely zero overlap between our respective fields of work. I'd never apply for a job with some Rails sweatshop run by a bunch of clueless hipsters. What a joke.
> From the perspective of the employer, Github, Bitbucket and other code sharing sites simply do not have these biases - when employers look at these sites, there is no demographic information on users to look at other than what the user voluntarily puts up.
Essentially, I can guess that any particular GitHub user is a white, upper-middle class man and be right most of the time.
I understand that this article isn't saying "We Use Only GitHub Profiles" when hiring but acting as though the code present there exists with zero social context seems willfully wrongheaded.
There is nothing whatsoever preventing a person of any sort from creating a code portfolio, and it's pretty easy to browse a github profile without learning anything about the demographics of the coder.
In contrast, basically every other screening procedure out there is potentially influenced by the demographics of an applicant. Prior employers/educators might have been racist/sexist, interviews are potentially subject to similar biases, etc.
> There is nothing whatsoever preventing a person of any sort from creating a code portfolio
You're wrong. Here are some likely hypotheticals:
* I'm lower middle class with four kids and don't own a computer that isn't owned by work. They own all the code I write on that machine and I'd be fired if they found out I was using it recreationally.
* My security clearance prevents me from publicly contributing for fear of getting swept in a security audit and losing my job.
* I've been harassed in the past by dogged pursuers who have doxed me and SWATed me. Publishing anything under my name is a no-go. If I don't use my real name, how do I participate on mailing lists and collaborate with other users?
So, yeah, I choose not to contribute and I own those choices. But pretending like there's "nothing whatsoever preventing a person" from open source contribution is, like I said, willfully wrongheaded.
That's the point of this article and Ashe Dryden's and the others linked before: there is demographic information associated with open source work in that a particular demographic overwhelmingly is able to contribute vs. others. They're not making idle suppositions; they're backing this stuff up.
From the perspective of someone who can make hiring decisions, how do you differentiate the two?
Besides, I see it the other way around: they're identifying "can't contribute" instead of "don't contribute" and asking people who make hiring decisions not to assume "don't".
Near as I can tell, all Ashe Dryden does is digs up some statistics about the overall population (women choose to be caregivers and earn less than men on average) and declares victory.
Among other things, if these statistics implied women can't create content and push it to free internet services, wouldn't that imply women couldn't contribute to facebook/instagram/etc?
As for distinguishing the two, I don't. My goal is to find good coders, simple as that. In the absence of information suggesting someone is good, I need to assume they aren't.
In the quest of finding "information suggesting someone is good", ignoring the known, recognized biases inherent in your sampling method seems like not such a good idea. You're throwing away data and then saying you have an absence of information.
Additionally, her statistics are not solely about women. They are about the multiple, separate groups that encounter obstacles when trying to contribute to open source projects. They include not just gender but race and health.
You're brushing away a lot of the subtlety of her argument and then misrepresenting her stance as "victory". On the contrary, she's exhorting us to work together to make our field more equitable and to recognize the deeper social realities underneath the shallow veneer of meritocracy.
Two things about your comparison:
1. There's a huge difference between contributing to Instagram/Facebook and contributing to GitHub that I'm sure I don't have to explain to you.
2. Dryden's blog post is not about "creating content and pushing it to free services". It is specifically about contributions to OSS not being inherently meritocratic.
>There is nothing whatsoever preventing a person of any sort from creating a code portfolio.
When I was hired by a former employer as a term of my employment I had to sign a contract saying all the code I wrote why employed for them, they owned. Even if I wrote it on my own time.
This has nothing to do with me anyway. I gave an example of a factor preventing someone from publishing code since you claimed "absolutely none exists."
> There is nothing whatsoever preventing a person of any sort from creating a code portfolio, and it's pretty easy to browse a github profile without learning anything about the demographics of the coder.
Not if they use their name on the profile or use a picture of themselves for their avatar.
You can stop reading after the picture, which happens to be at the top.
Why doesn't every programmer have a great GitHub profile? The same reason not every mother of 3 children looks like a fitness instructor. Priorities.
You write C++ all day at a bank, can't use any open source, and your mind is filled with the intractable legacy issues, politics, etc. Maybe you're married and your spouse wants to go salsa dancing on Tuesdays. Should you prioritize throwing some Ruby into GitHub? If you want to work for a startup, yes, you should. If you don't care, don't bother.
I find that this whole Github portfolio thing is getting blown out of proportions. A Github profile is a useful extra, but shouldn't be the only indicator of programming capability.
Neither should education and previous experience.
When hiring, people should talk to each other about something they've done (be it public code or private code). That should give a first impression.
Then you either hire or not on that first impression, and use the following 6 months re-evaluating your first impression. That's what it's there for.
If either side is not fully content, then's the time to decide to keep or fire.
> Perhaps the candidate has little experience because previous employers he interviewed at did not wish to hire a Black Muslim Lesbian T____ due to fears of getting sued.
First of all, I'm confused about how someone can be a lesbian and take 'he' pronouns. Second, 'black', 'muslim', and 'lesbian' are names for identities, while 't____' is a slur. Its presence in your post is incredibly jarring and I nearly stopped reading there.
Woah I blew past that line in the original document. This guy is a jerk. Using language like that is unprofessional, and basically asking for trouble down the line.
Thank goodness due to his hiring practice he'll screen out all the candidates who will be remotely different from him, so he can keep on living in his bigoted bubble!
FWIW, the correct term is transgendered, but most transgendered people I know are not offended by "tranny." It's all about the context. "That fucking tranny" will be offensive while "I love my tranny, that bitch is awesome!" will be looked upon as a nice thing to say.
Regardless, the author is trying to use shock value with the photo at the top and by including conversation that is entirely irrelevant.
I can change the language, though I don't know what the right term is. Transsexual?
I've only met one person of that category in my life, and she referred to herself by the same term used in the post. What's the official term to use? I don't want this distracting from the actual point being made.
Nice to start off the morning with a laugh. I'm not interested in the boring conversation about whether trannie is an acceptable term to our society, but the idea you had no idea what it is short for is a joke. As a graduate in Sociology, how can you understand "how technology shapes culture" and obviously never have bothered to know the name for large subsets of our culture?
I wrote most of this, but the name on the page is my cofounder. I just forgot to create a tagline for myself before posting it. My background is math/physics/trading.
That's not an excuse. You seem to have an ignorance about a whole group of people. I'm fairly sure even the Pope could come up with the word "transsexual".
I don't mean to imply that ignorance implies anything else about you, but it's ignorance. You can alleviate it by:
1. Meeting some people outside of your apparently somewhat homogeneous bubble
2. Reading books
I'm not making excuses, I'm pointing blame at the right person. (The tagline describes my cofounder or a fictional person, probabilistically. Haven't updated it yet. )
He's making a point that candidates who are of a particular class are often screened out of employment based on that class. He's not saying he does that, he's using hyperbole to illustrate a point. The very usage of the incorrect pronoun likely emphasizes the hyperbole even more.
"Just please don’t send us proprietary code that you have written because we don’t want to be sued by your former company."
This is the gist of the problem. The last time I did any serious development unencumbered by IP constraints was 15 years ago. Anything indicative of what I write professionally, I can't send you.
If you are fine with ruling me out of your hiring process (and everyone in a similar situation to me) that is fine. When I'm hiring, I don't like to restrict my pipeline that much.
Maybe this is my own biases showing, but it seems like the main proponents of the Github hiring practice are folks who don't have a lot of experience writing software. For inexperienced developers, I can think of nothing better than a github profile to prove your worth, but nearly everyone who has written a lot of software will have been encumbered by IP agreements at some point in their career and I'm never surprised when someone says they can't show/talk about something they've worked on.
Stopped reading at the unnecessary photo. Sorry, this is getting stupid.
On topic though, I didn't see where he linked to his "portfolio" or his company's body of OSS work that these OSS rockstars he's planning on hiring will be able to continue the great work they have been doing. Seems to be all talk no action here.
I think most of us can agree that FOSS contribution is relevant and interesting and we'd certainly pay it due notice. I would always consider it, check it out and even in some cases expect it from any senior hire. However, I don't think this is a very groundbreaking or controversial idea on it's own. The idea that it's mandatory on the other hand... well honestly the entire post is click-bait crap so way comment on that thought any further.
This blog post really misses the point of the previous discussions about using Github as a recruiting screening tool. It's fine to use it as a small 'plus' factor, but to screen candidates who have no public code portfolio out is silly, counterproductive, and discriminatory.
Also, for a guy who graduated with a degree in Sociology he sure does talk like an asshole.
Well, of course it's discriminatory. The purpose of a hiring process is to discriminate in favour of the kind of people you want to hire and against the kind of people that you don't.
I don't think this article really misses the point of the previous discussions so much as it argues that the issues raised in the earlier discussions are outweighed by the usefulness of a code portfolio as part of a hiring process. That's still debatable but it's something that reasonable people can disagree on.
I believe the parent was trying to say that it's potentially illegally discriminatory in the US due to the fact that FOSS is gender and racially biased hugely in favor of white male participation. If you use it is a screening tool to cull candidates, there's probably a case for you breaking US hiring laws.
That strikes me as unlikely. I might just be naive, but I can't imagine being sued for filtering candidates on their code portfolios.
This is a genuine and non-rhetorical question: how does it differ from academic qualifications? Both are unpaid work that a person undertakes [partly] in order to prove their value to potential employers, but nobody is getting sued for filtering candidates based on academic qualifications.
This relatively new "show me your github or I'm not interested" fad is getting old. If you are so narrow minded to judge development proficiency based on having an open source portfolio, then sorry I'm not even interested in working for you. You are more than likely going to lose some talented developers with this approach. I get the fact you want to see some code, there are other ways though than taking this hardliner approach. Not two days ago I ran across a job posting on SE that was very clear that if you did not maintain a public github profile they weren't interested at all, just shocking to me.
My Excuse? My previous employment I coded in .net and php for nearly 5 years as the only developer at the company. Personally wrote thousands of lines of proprietary business code I simply can't show you. Routinely put in a lot of time in and outside of work. I was coding and committing code before GitHub was cool, hell, before it was even founded. I work on my own now getting a few ideas off the ground and doing freelancing. I just checked my local SVN repo, I have about 25 projects in there over a few years (some massive). I routinely work 12 hours a day for clients and moving my projects forward. None of this code is public.
1. I don't use github, nothing again git or github at all. SVN works well for me, fits my use case and I know it well. I know how to use git, I have a github account with a few small contribs to javascript libraries I've help extend, but that is all. It is simply .01% of a reflection of the code I've written.
2. I would love to contribute to open source more, but I have priorities, life, family, my own goals. There simply isn't enough hours in the day.
3. I could talk ALL DAY LONG about my experiences coding various projects using whatever framework you want to talk about. I can show you code for my personal projects, although it isn't public. However, if you want to judge me on my lack of github/bitbucket, then so be it.
Last week I worked about 80 hours (salaried, large co.). I have a wife and three kids. Any time not spent at work is spent with kids or chores, usually simultaneously. GitHub is somewhere south of these priorities. If you as a potential employer are not ok with this, that is certainly your prerogative; I would even agree that spending that amount of time at work is not ideal.
Firstly I don't think anyone really needed a 1000 word essay on why hiring is inherently probabilistic, I think that much is obvious.
The ultimate issue surrounding using github as signal is whether the presence of a github profile is inherently meaningful signal independent of every other factor. People who advocate github are usually implying (either explicitly or implicitly) that someone who codes outside of their day job (and then uploads it to share with the world) is more "passionate" and therefore a better coder. Many people disagree with this. It's been repeated ad nauseum why this reasoning is faulty: everything from getting your fill of creative work during your 9-5 to maintaining good work/life balance. Just the presence of a github with some activity is not an indication of passion or more importantly quality code.
Of course, people who defend using github like to say well I can know a candidate better by looking at their code samples before hiring them. Again, this absolutely goes without saying. If you want a code sample, just ask for one. I'm sure most people can provide something for you to peruse. If not, keep a simple task handy that you can give them to crank out in a few hours. Such an easy solution to this non-problem.
The reality is this is all just the rationalization here. What's really going on is people use github because that's what they're into. "It's obvious that I'm a top developer, therefore my behaviors are indications of top developers". Bullshit. This is just another symptom of the "hiring in my own image" phenomenon that's rampant in this world. Those of us who claim to be rational must recognize this and realize it's nothing but narcissism.
And no, the claim that anyone who is against using github as signal are people who aren't passionate/would be rejected/etc. I have plenty of code that I write on my own to scratch particular itches. Everything from machine learning scripts to greasemonkey scripts to patches to borked java applications. None of them were written with wider consumption in mind, and so I am not prepared to release them to the world. If you want to see it I have no problem showing it to you with the understanding that it may not be polished and not always follow best-practices when working on a team. Anything I produce is a reflection of myself so anything I release to the world must have a certain level of effort and polish put into it, otherwise I simply would not be comfortable.
If you do the coding test before interviewing the candidate, you are wasting the time of many candidates who will be screened out for being insufficiently hipster, having Windows on their resume, and similar non-coding related reasons.
Is he being cheeky? Why would you screen someone out for not being a hipster or having Windows on their resume?
I've gone through the whole hiring rigamarole, done the coding test, etc, and been rejected at the last minute due to a lack of culture fit. (And justifiably so.)
The point is that I'd rather have that happen after a 20 minute phone interview than after a 2 hour code test.
Humm. First, this blog post was not a quick read while taking a quick break to read a hackernews post, so read if you have time. Second, while being able to see what a developer's programming style and their interests are like via a code repository, not having a code repository does not mean someone is not a good programmer worth hiring. It seems like those who will ONLY hire someone if they have an amazing github are asking for people who do NOTHING but breath, eat, and sleep code; someone without a life, someone who does not cook, have pets, loves someone, or has other interests than being in front of a computer uploading code to github. When you leave home at 7AM to be be at work at 8AM, and work until 5:30PM, and then you get home to cook dinner, feed your pets, clean a little, MAYBE get to the gym, it's already 11PM or midnight. When exactly do these employers think someone has time to do stuff other than work and take care of real-life responsibilities? I mean, yes, programmers could use the weekend and some if not most do, but if moving to another job means having an amazing code repository, this would not be built over one weekend, maybe not even several. There are brilliant programmers out there who simply do NOT have TIME to spend it creating code for employers to see on a code repository site, so requiring one is in fact discriminatory to those programmers who just cannot spend the time adding to the code repository.
I have personally seen how employers think they are so awesome that they will not even consider a programmer for hiring because of lack of a github. All I can see say is that in the case I have seen, all those employers missed out on hiring a great programmer, one who is truly dedicated to work and programming, one who really does live programming while still having and enjoying life.
While "[l]ooking at a code portfolio for people who program is fundamentally no different from looking at past experience, at educational credentials, or at any of the other standard resume screening techniques..." may be true, I think that it sidesteps another important consideration: What does requiring a code portfolio signal to potential applicants about the employer? Does it signal that the employer will be inconsiderate of those who can't share their professional work, or those who don't have lots of spare time for side projects? Does it signal that an employer is more interested in the placement of curly-braces than the ability to communicate verbally? Does it signal that the employer is solely interested in getting tomorrow's work done, or do they balance increasing the value of their codebase with increasing the value of their employees?
Maria Kang makes no excuses, unlike developers without github profiles.
It's pretty ironic you used this picture, since she got a lot of negative press for it, and she mostly used it to promote herself and her blog.
Women are already under enough pressure in society to look "perfect" and this picture shamed them into not looking like the model. So women software engineer are supposed to spend tons of time at the gym to look like the model. They already do most of the child rearing (as well as house keeping and errands), and on top of that in order to get a job they are required to code in their spare time just to get a job.
I live pretty far away from my parents (about 200 miles), and my dad is in poor health. He almost died a few months ago. I spend a lot of weekends visiting because I don't think I have much time left with him. Yes, that is my choice.
Factors outside of my control prevent me from moving anywhere else.
There are tons of potentially good candidates you would miss by rejecting per default when there is no public code to see. Just look into the comments here.
The group I want to point out.
Very experienced programmers who have been in the business more than a decade who get paid for doing what they think is interesting and challenging because they are so good so that they can choose their job and still get paid. Most of the time they throw themselves into work and their current project. Why on earth would they want to invest more time into a little pet project which is probably ten times less interesting than their current job?
From my perspective a GitHub portfolio might not be required to get hired, but if the code you've shared is bad code then it could keep you from getting hired.
Edit: That doesn't mean I wouldn't have one. I do have one. I just imagine that a bunch of developers sitting around judging someone on publicly available code may not always be considered a positive.
BaynesianWitch - The company that used the word 'tranny' in an A/B tested public statement intended to incite and offend their audience. Stay classy, BaynesianWitch.
From the blog, it has the following piece of code:
def outcome(hiring_procedure):
allowable_candidates = [c for c in all_candidates if hiring_procedure(c)]
candidate_values = [risk_profile(value(c)) for c in allowable_candidates]
return candidate_values.sum() / len(candidate_values)
What happens if we get to the return line, with an empty candidate_values? ZeroDivisionError ?
The short answer is, that code is a lie. It should actually have been the integral of risk_profile(value(c)) dP(c).
But the python code is close enough to illustrate the point, and I've found people are far more likely to read and understand python code than equations.
This part of a broader problem with the industry. Tech companies subscribe to stupid hiring fads (see Google's earlier reliance on brain teasers) and apply hyper-specific screening criteria (10+ years of .NET experience), then bitch and moan when they have winnowed their applicant pool to nothing and can't meet their recruitment goals.