Hacker News new | comments | ask | show | jobs | submit login
I have no side code projects to show you (codementor.io)
771 points by nnd on Oct 25, 2017 | hide | past | web | favorite | 522 comments

I’m not mad at you, but as the dude on the other side of the desk, I have to decide whether you can cut it and I’m fine saying no. I could make you do some whiteboard problems, but I think whiteboard problems are pretty far removed from your day to day development and I am uncomfortable relying on them as a proxy for your ability to ship code.

I don’t really disagree with your philosophy of working while you’re working and NOT working in your downtime. I think the same way.

But I have a whole pile of stuff I can show you that I’ve worked on. I don’t ask anybody to take my word for it. In computing, you just can’t. The signal to noise ratio is too bad, there are too many guys out there making unmaintainable junk, if they actually ship anything at all. There are shops out there that can afford to hire a bunch of maybes and just weed out the duds when they wash out, but I don’t have that luxury. Nor am I going to go on some weeks long hiring safari where I hire you for a week or two and see if you can cut it. No time man. Sorry, don’t have it. I’m going to give you the job or I’m not going to give you the job.

On the flip side? As somebody who has plenty of open source projects--the set of employers who actually looks at them is within epsilon of zero. I have not once, not once gotten a meaningful question from anything I have posted on Github in any interview or client meeting even after suggesting up-front that it's worth taking a look. And I used to interview on a monthly basis just to keep in practice! No--instead you get dance-like-a-monkey "coding tests" (unpaid, of course) that are nonportable from employer to employer.

Employers don't care. Employers say they care, but it's signaling. If I didn't write code on my own because I like it, it would be a serious negative to do so on the off chance that you run into the black-swan employer who takes the question seriously.

Hiring manager here who reads Github profiles -- it just occurred to me that I've almost never actually mentioned it to the interviewee. It's flavoured the conversation for sure (except for the bootcamp grads who all have identical Githubs) but I tend to talk more about your process than the actual code.

When I've had time to prepare, I've taken the time to review candidates' GitHub repositories, and it has surely helped to really understand what "open source experience" means vs "coding camp attendee". It's fun to ask about a particular class or set of files and ask the candidate to teach me something about the project. It shows we care, and also helps me tease apart a candidate's communication abilities.

As a software engineer who does technical interviews, I also regularly review GitHub contributions if mentioned on the resume. It's not a deal killer if someone doesn't have these things...but if you're up against another candidate who does (and their work is good)...then the other candidate is likely getting the job.

When I hired people, I looked around on the internet for people with public work (whether side projects, university homework, open source development, essays, whatever), I scrutinized all of it, and then I talked to the people about those things.

Because I myself feel that this is such a good and reasonable way of hiring people, I wouldn't want to work for an employer that wouldn't let me do open source work.

I mean, it's also because I consider open source to be the public sphere, and I really do believe in software freedom. Maybe my professional metaphor would be the coder as a writer. Of course I want to see your published work.

What % of candidates have side projects to show you? When I did interviews, I saw probably less than 10% have any side projects to show.

I work in consulting, almost entirely for federal and state government. Here that number is low single digits.

This is what I think most people fail to understand. It's not a _requirement_. The vast majority of people DON'T have side projects. Companies can't afford to simply auto-reject everyone who doesn't have side projects. The point is that if you do have any, it can help make you stand out dramatically.

That's a surprising number. Did you interview mostly recent grads?

Not surprising at all. None of my colleagues have anything public. If that's not your thing I struggle to understand why should you.

And most of my work mates have public open source code.

I would guess the environment candidates are selected from influences these numbers significantly.

And you're right, I'd understand that in some of these environments, OSS isn't necessarily considered fun or relevant.

Then again it's all speculation.

> except for the bootcamp grads who all have identical Githubs

Out of curiosity, what do you mean by this? What sorts of things do they all have in common?

From my experience (of talking to bootcamp applicants), bootcamps are not so much about learning how to code, but rather how to get an entry-level job. At first glance of somebody's Github, it will look like they have lots of open-source projects and collaborate with other people. But this is just part of the bootcamp "script".

Bingo! Bootcamps are basically multiple month training sessions for interviews. Princeton Review for dev jobs.

Regardless, I actually think this paradigm works for entry level jobs at most companies. You don't have to know anything for your first year in most cases. You just have to be able to show up and follow instructions.

Basic Angular project. No contributions. Poor code. CS 101-style projects.

How deep of a reading is your reading of the profile though?

hope you guys realize there's more to opensource than github..

show me freebsd or linux kernel contribution on github, for example

Here you go: https://github.com/torvalds/linux/commit/692b48258dda7c302e7...

I do understand what you're saying though but I don't think people are going to just search Github for your name on the off-chance that you're on there, I suspect that what most people do is look at the things that you've specifically mentioned in your CV.

Searching Github for your name can backfire. My name is not very common, but there are 2 people on Github that have it and I'm not one of them. I do have a huge body of stuff on StackOverflow though.

> Employers don't care.

The wrong empoyers don't care. The right ones do. My github portfolio is a means for me to determine if a potential employer is suitable for me, not the other way around. It's not a test for me to pass, it's a test for a potential employer to pass.

If a recruiter brings up things listed on my LinkedIn page, I know they're not suitable for me. If they bring up things on my github, I'm reasonably positive I want to talk to them. One company's CEO asked me a whole bunch of questions about one of my projects. Not as an interview, but as a peer. He had a genuine interest in my side projects. If he'd asked me to come work for him, I would have said yes in an instance.

I don't expect my surgeon to do surgery in his spare time. Nor my plumber, dentist, gardener, or any other professional. I look at their past work history and reviews from customers and/or coworkers.

I will never understand the belief that side projects are in any way necessary. I'd much rather have a well rounded individual who isn't moonlighting on other things but rather can bring his clear focused mind to the office.

> I don't expect my surgeon to do surgery in his spare time.

I really like it when my specialist has a prominent public profile.

A recent case: tenure at a major university, had written or was quoted in many good journals and papers, and had a number of impressive accreditations (for example he was also head of his specialty's medical association for the country). This showed dedication and interest as well as some measure of intelligence.

I was accordingly confident he was on top of the latest developments in his field and would be able to deal with complications better than your run of the mill specialist. The care was great throughout.

Similarly, our lawyer has been quoted twice by the Supreme Court in Singapore, has written many articles in the Straits Times, and does great (and visible) pro bono work, all of which I could find off a quick Google for his name. Whilst I hired him based on a recommendation, it was comforting to see this public confirmation of his ability and interest in the field.

While these points may well be examples of this person's ability, certain credentials don't have to mean as much. As an example, the head of an association is not necessarily chosen for their merit.

I knew of at least one department Dean at a college that was chosen for that position so he would not teach anymore students.

Not so sure about this. It would be OK for my surgeon to just be a badass surgeon and not a politician. Public confirmation doesn't add much because the skill of navigating that is (mostly) unrelated to the skill of achieving a great surgery outcome.

I think the key to your answer is in the parent's text: "This showed dedication and interest as well as some measure of intelligence."

My reasoning is that if someone is very good in his line of work, some of that work starts to spill out to the public, either by research, teaching, articles, blog posts, etc.. If a person is not a pure introvert, working all by himself, and is quite good, invariably someone will notice it..

Me either. It's become cool to work 60 hours a week because you like your job, and then go home and do more programming because you like programming.

Why it's cool... I don't know. I agree it helped me in the early days of my career, but at this point, as a senior developer I can pick up new technology much faster than my early days. I can come up with a working solutions much faster.

I like my job, and I like programming... But there are other things I like too.

So, why do I need to work long hours and program outside of my job? I have good enough communication skills to convey my experience. If I'm not hired for not having an active github account, or a side project, there are five other companies that would.

I like my job, and I like programming... But there are other things I like too. Such as playing the guitar and saxophone, writing side coding projects, mountain climbing and traveling. Do you realize that writing side coding projects is just one of the many things we like to do?

Just because we do side projects does not mean that we do not have a life outside of coding.

I agree with you. Everyone has their own hobbies.

But certain people believe that in order to be a good coder, one of your hobbies MUST be coding.

One difference is that many "programmers" actually don't do any programming at work. Their hands on experience is a 5 week java course in uni 10 years ago and after thay they have been stuck in meetings or writing brain numbing CRUD in a proprietary domain specific scripting language at $bigco ever since. The only way for many to get actual hands on is to do side projects.

Hair dressers don't have this problem, they are actually cutting hair 8 hours a day.

Carpenters have a more hands on education compared to compsci which is more maths than version control.

Doctors require 2 years of practice supervised by other doctors before they are certified.

There are a lot of more senior doctors in admin or specialty positions who aren't making rounds or doing surgery anymore, however.

yeah but they did a lot of medical work before getting there though

> I'd much rather have a well rounded individual who isn't moonlighting on other things but rather can bring his clear focused mind to the office.

I don't understand how you equate doing side code projects to not bringing clear focused mind to the office.

Would you also much rather bring a well rounded individual who does not take art classes after work? Do you think taking arts classes after work would affect one's ability to focus in the office?

Then why do you make similar assumptions about doing side code projects? For many of us, doing side code project is like taking art classes. It is art for us and it helps us learn new things about the art of software development that is not always easy to learn in the constraints of real world software business. Most importantly, it is a relaxing and enjoyable experience. It refreshes our soul and combined with a good night's sleep, it helps us to better focus next day at work, not the other way around.

Are you saying that you would much rather not hire people like us who care about the art of software development because from your high horse, we don't look like "well rounded" individuals? If you hold this kind of irrational conservative beliefs, I'd much rather not work for you.

Thanks for pointing this out. Working on my humble side projects is the hiking/cooking/traveling of my life. I spend my free time programming for the same reason the author doesn't spend his free time programming: because life is too short not to do what you love when you can. Making progress on these projects helps me to be more present in other aspects of my life, like my job and my marriage.

If the person enjoys their side projects, go for it. I'm not saying they can't. But on the hiring side I'm not going to require that all employees have side projects that match their day job.

> I'd much rather have a well rounded individual who isn't moonlighting on other things but rather can bring his clear focused mind to the office.

I don't think it's wrong to work on side project. Some people like or want to. Who am I to judge what they do on their free time? It might happen to overlap with their job, or it might not.

But I agree, I don't require my surgeon to do surgery in his spare time. I've worked with plenty of great engineers who wanted to leave work at work, and after 40 hours of engineering, enough was enough and spent their free time doing something that wasn't engineering. I've also worked with people who did side projects, and were equally amazing.

A lot of coders consider coding not only as a job, but also as hobby. Usually they experiment with new stuffs this way, and thus their open source projects at Sourceforge/Github, etc etc.

I don't expect a brain surgeon to mess with somebody's head because he's bored at weekend, though :p

For the first many years, I was this way (considering coding as a job and a hobby). I guess it still is in many ways, but as I've gotten older I've started to feel the burnout more. If I spend my weekend doing anything on my computer, I feel it on Monday and just want to do something different.

Personally, I've lately been considering the opposite: getting out of coding as a job, so that I can continue to love it as a hobby.

Yeah, I tried that. Turns out I missed coding at work a lot and will now go back to it.

You're right, I was probably swinging too hard on the other direction with my comments. The core of my argument stands: I refuse to require side projects as a condition for hiring someone.

Surgeons, plumbers, dentists, and gardeners are licensed, bonded, and insured, entrusting the certification process to a third party. Surgeons and dentists also have professional development baked into their work schedules, which software engineers should have -- say, hire someone for 35 hours per week so they can spend 5 hours per week on professional development.

The other difference is Surgeons, Plummers, Dentists and even Gardeners get payed for services rendered not time.

This is a huge problem in the industry and is the source of massive waste.

You have a rather rosy view of medicine and gardening. Your view of plumbing and dentistry aligns with my own, however.

I would expect my doctor to read papers, news, etc to keep abreast of the latest medical advances though.

This! And similarly, software developers should keep abreast of latest technical advances.

What makes software engineering different from surgery is that when a surgeon learns something new, he does not have a live patient at his home to practice the new thing he has learnt, but as software engineers, most of the times we can immediately fire up an editor and write a PoC or a toy project to practice the new thing we have learnt.

Its strange that such natural and obvious curiosity and desire to try something out outside of work deserves the kind of criticism that it is receiving in the original article and some of the comments here.

> Its strange that such natural and obvious curiosity and desire to try something out outside of work deserves the kind of criticism that it is receiving in the original article and some of the comments here.

The critic points out that it is no longer curiosity which drives you when people expect you to do it.

Yes, it is normal to tinker with new stuff to get acquainted with it. It is, however, not the same thing as publishing results of that tinkering on e.g. github, or blogging about it as if you want to break some posts-per-month record.

I do not get this obsession with putting everything out there. I have a huge local directory with my old test programs, PoCs, started but abandoned ideas, etc. I regularly try new stuff, new libraries, programming languages, what have you. But my github account is empty.

The internet needs less stuff on it, not more.

> I do not get this obsession with putting everything out there.

It's not obsession. It's just a matter of convenience. git add; git commit; git push; and all my experimental code is safe on someone else's server that is managed much better than I can manage my set of USB drives and hard disks.

> The internet needs less stuff on it, not more.

I disagree!

I expect a surgeon to stay up to date on the latest in medical technology. And I don't know too many surgeons who clock in at 9am and clock out at 5pm. Many surgeons pull 60-80 hour weeks regularly.

> I'd much rather have a well rounded individual who isn't moonlighting on other things but rather can bring his clear focused mind to the office.

What I do after the office hours is really none of your business. I should be able to do whatever I please (write code, play guitar, etc.) after office hours without you raising concerns about it.

If you believe I do not bring my clear focused mind to the office, then that is a separate issue to be discussed separately without bringing in what I do after office hours into the discussion. It is my prerogative, not yours, to decide if I need to give up what I am doing after office hours in order to focus better at office.

People who have only recently completed their 10,000 hours don’t understand that there are diminishing returns, and that learning something else entirely may help you more than doubling down on your narrow skill set.

Some of the things you can learn outside of this field are what overtraining is (actually not that different from the AI concept), perspective on the learning and teaching process, the importance of preparation and good tools, patience with new ideas, and the dangers of trying to rush through the learning process.

Besides the fact that many consider the 10,000 hours a myth, please keep in mind that the theory is about 10,000 hours of practice.

I bet that even people who have spend 20 years on side projects haven't reached those 10K hours of practice yet.

We aren’t talking about people moonlighting as developers. We’re talking about people in the industry working on software every waking hour.

Roughly 220 work days a year, 5 hours a day at work if you are full time, 6 if you dabble a bit outside of office hours, that’s around 5 years out of school, which isn’t too far off from my experiences prior to Malcolm Gladwell.

Working != practice.

When I'm building some CRUD interface for my job, there's nothing I'm practicing with that. When I spend my entire day doing that, I've gained 0 relevant hours for Malcom's 10K.

No, side projects are not necessary. They're just a signal that you're really into this stuff. Like all signals, it can't be evaluated in isolation. It's no different than preferring candidates who come from a top University.

Some jobs are easier to do on the side than others. I'll bet the gardener has some side projects too.

If you were a surgeon looking to hire another surgeon for your surgery team, I bet you would want to observe her performing a surgery.

If you were looking for a drummer for a band, you would want to hear him play.

The former one, maybe not. I'm sure lots of practices hire surgeons by looking at their record and not observing them actually doing a surgery. After all, it's the results that matter in the end.

The latter one, sure. Makes me wonder if we should call interviews as auditions.

What about an artist or a musician? Would you expect a musician to practice and get better in their free time or only play their instrument during paid gigs?

Fine artists usually produce works on speculation. Craft persons usually produce on a cottage-industrial basis, essentially commodity piecework. There are a lot of ways to get paid, beyond the two you highlight, so the parallels are not as clear as you appear to be assuming.

A surgeon is licensed. Many times they will publish papers etc. especially if they are trying to get ahead.

You list things on linked in just so you can blame potential employers for looking at it? I dont even have linked in account and still find it ridiculous to hold it against them if they dare to look there.

“How much time is the interviewer willing to invest” is a super legit question to be asking. It tells you a lot about how many shits they give about you in particular and is a good proxy for the quality of the team in general.

It's not that they "dare to look there", it's that they _only_ look there and don't care enough to look even a tiny bit further.

> My github portfolio is a means for me to determine if a potential employer is suitable for me, not the other way around. It's not a test for me to pass, it's a test for a potential employer to pass.

Nice! Agree with that.

Thats a pretty arbitrary line. Why wouldn't they want to ask about hte projects you poured thousands of hours of paid time into?

I disagree with you, and grandparent, and the article :-)

I've hired plenty of great engineers who didn't have side projects on github. I have never specifically been looking for people that don't have a life outside the field, and unlike GP I have found that there are other ways to evaluate people that can identify great engineers with very few mistakes.

But that doesn't mean I wouldn't see your high quality or impressive open source code as a huge plus. Quantity doesn't matter -- if anything if you're super prolific I might wonder if you have much energy for your day job -- but I'm not going to pass up a chance to see what kind of work you do.

And because quantity doesn't matter, I think that writing a little bit of code "for your resume" (every once in a while, to reflect your continuing development as an engineer) is a pretty reasonable investment even if you don't happen to want to spend your whole life outside work programming. It's not required, but it might get you a really great position some day, even if like the OP you're up front about having a life outside of software engineering.

This. Not once has a potential employer ever said to me: "Based on your Github, you can obviously fizzbuzz with the best of them, and know a for-loop from an if-statement. Let's skip the phone screen/homework assignment and get you onsite directly."

Speaking of FizzBuzz, I had never heard of the fizzbuzz test until recently in interview. In 15 years of frontend work I've never needed to find if a number is divisible by another number, so I forgot about the modulus operator '%', and stumbled on that question. Apparently not knowing about that made me unqualified for the job, above all other evidence of my frontend work and achievements. It takes less than 10 seconds to find the solution, but in interviews we are expected to have the entire JS reference guide in memory.

There are plenty of ways to write fizzbuzz without the modulus operator, although it is the most obvious way. But you can also combine infinite streams, or keep track of the index of the next "fizz" and "buzz" with simple addition...

Perhaps there are plenty of ways, but finding the remainder of the division is what's needed and in my case the test was pen and paper, with just a small space for the answer and no way to test my code.

It's all good, it wasn't the role for me, I'm more creative dev than programmer and I didn't get a good impression of the company anyway. I've moved on.

Maybe it's better to provide access to Google on the interview. What do you think?

Or at least the JS reference. Or design the questions to fit the role. If it's a frontend role, then ask frontend questions, not some bullshit high school arithmetic.

I wrote most of the 'fizzbuzz' function but couldn't remember the "%" operator. Tests are annoying because it's not how developers work, at least I don't work that way, without any way to verify syntax or look anything up.

I recall an interview from years ago where I was asked to show the code behind my websites and step through it explaining what it did and why I did it. This is better than a test, because you're explaining how the code works and also showing off your work.

Time for anecdotes. Twice in my career, an employer has looked at my GitHub and asked specific questions about certain projects, what motivated me to write the project, why did I not use a popular big data technology to write the project, etc.

But, did that interest lead directly to a job offer? My guess would be no, knowing the industry and given no other information about your situation.

There was a time, last year, when I'd regularly receive email-SPAM from recruiters talking about one of my C++ projects. (Console based mail-client with Lua scripting.)

Later I moved that project to a github-organization, as it didn't need to be tied solely to me, and I wanted to allow others to commit, etc.

I'm not sure if the SPAM-mails went away because recruiters stopped scanning github for profiles, or because there was no obvious link between that particular project and my personal email.

It doesn't happen too often - but I've had that happen several times, actually. Just like people might pretty much hire you (or invite you to substantially greenlighted interview process - sans whiteboarding or over-the-phone grilling) solely on the basis of a talk you gave, or an impromptu discussion about some topic you're insanely interested in, after meeting you at some event.

Granted, this has been for gigs (or temp-to-perm) situations, not official, up-front fulltime roles. And some of these people ended up being flakey or sketchy (as spontaneous / whimsical people sometimes are). But nonetheless, hired (or substantially fast-tracked into their process, bypassing most of the usual interview crap) - pretty much on the basis of single demo or code sample (or in one case, literally on the basis of a single stats plot).

Because by this point, pretty much everyone knows how silly those live fizzbuzz sesions are anyway - including those administering them.

That's only because most recruiters are non-technical. They lack the ability to gauge your ability.

> As somebody who has plenty of open source projects--the set of employers who actually looks at them is within epsilon of zero. I have not once, not once gotten a meaningful question from anything I have posted on Github

I guess that isn't the idea here. I think the idea is for them to grab a dev from the next cubicle/floor, point them to the repo, ask "can he code?", and the dev says "sure looks like code to me, yeah."

"Showing up" is 90% of every move. Sure, the guy who copy-pasted something together could get the job that more talented dude who was too timid to share their stuff would be much more suitable for, but such are the trade-offs here.

In my experience, that doesn't happen. And I go on a lot of interviews; it's an activity that helps with a set of skills that I always need to keep sharp. And, along those lines, I get a lot of offers! But it has never-not-once been related to my Github profile.

In my experience, it is employers signaling how cool they are before they ask you to do their multi-hour coding test for no money.

> before they ask you to do their multi-hour coding test for no money

Where do all of you keep finding those companies? Is that an SV thing, or am I just not applying to "prestigious" enough companies? I've been on my fair share of interviews and haven't had that happen once, along with no whiteboard interviews. I can echo your experience about Github projects playing a small role apart from signaling though. Once, an interviewer mentioned one of the Github projects I worked on and we exchanged 2-3 sentences about it, but I don't think it had a significant impact on the result of the interview.

It's pretty much a standard interview practice in Boston.

As a rule, I don't and have never consented to them because I find them exploitative.

They're pretty rampant in Canada. You usually get one of these types:

- Long home coding challenge (for no pay of course) - Hours of whiteboard coding talk. If you're lucky it relates to the job at hand, most of the time it doesn't - Doing coding tests online, either through their propitiatory app or through sites like HackerRank. - A conversation type interview, where you talk about the job (have only had this one in my recent attempt. Didn't get past that phase)

Of the roughly 30 interviews that I've done in the last ten weeks, only one place ever asked about my github profile (which I don't have yet, currently revamping it since most of the code was years old and pretty meh). Almost every one of them had questions about my linkedin profile though.

>>Long home coding challenge (for no pay of course)

There is a way to do this. But some people go a little too far.

Last years one start up asked me to write an entire backend application for them for free.

Sounds like a large retailer that wanted the same for an interview. The job I found out later went to a kid of one of the managers.

That they declared bankruptcy this year has been a tasty bit of schadenfreude.

Just write one that will explode in their face within two or three weeks, and remove that time bomb when they hire you. :)

Can confirm, not a SV thing, plenty of it in Europe.

Nothing to do with prestige, it's a simple "will this monkey jump through hoops" test to see if you'll work in shitty conditions for shitty pay.

My go to response is asking them to look at my git instead, which ends the process in about half the cases. The other half they look at it for <5 minutes and point out the error that isn't an error (yes I know what an SQL injection is, and I've taken steps to make sure it won't be an issue).

If I came onto your project I'd be trying to fix that SQL injection too. Why don't you add a comment explaining why you know it's safe and why you aren't using the standard way of escaping variables in SQL?

That actually didn't occur to me.

I'm not sure where I'd put the comment, since in the part where router paths are set up, it's self-documenting/obvious that it isn't an issue, and if the person skips that and looks at individual functions only, then I'd have to put it in too many places.

I guess I'd put it in the paths, as it's something that needs to be kept in mind while updating the paths. Also in the coding conventions, but it's understandable not to have one on a small project.

Another option is to create a function like _escape_numeric_string that returns its input after asserting the string is numeric and has a comment explaining, and use it everywhere. Although yes, that does mean changing all your individual functions.

Generally if a code reviewer suggests a change it is worth documenting why the change isn't practical. That can be adding a test or adding a comment.

I recently had 3-4 interviews in the same format. Its so frustrating and heart wrenching after rejection. For a job i have to go through 2hrs of coding interview.

I've had to jump through those terrible "online coding challenge" hoops for jobs that didn't even involve coding! Tech companies far and wide are currently obsessed with them, to their detriment.

No--instead you get dance-like-a-monkey "coding tests" (unpaid, of course) that are nonportable from employer to employer.

Anecdote time! I had a potential employer ask me to do the tutorial for their framework before the interview. I submitted a pull request to fix errors in their documentation during the interview. I got the offer but didn't wind up taking it. I don't do those types of things anymore because it's more beneficial to them than it is to me. I could see the value in doing it if I was a junior dev trying to get my foot in the door though.

I had something similar where they had a test app online, but the test environment was broken to the point of it being non usable. Emailed them as to what was broken including error messages. Got an email cancelling my interview twenty minutes later.

Hopefully they weren’t trying to hire you for a QA position...

From my brief interaction, I';d be amazed if they were aware of the concept of QA

I disagree, my side projects have led to many employment opportunities, including my current one. I also get recruiters contacting me and mentioning my side projects. I often suspect their system just found my projects and dumped them into the email, but other times the email is clearly from a human who has read my blog and/or investigated my projects.

I've a bunch of personal projects on github.

They are not that successful, a few dozens stars generally for the most serious ones. And also, there are not especially trendy (no fancy react app or js library, but stuff in python (lib, web ui), C, a puppet module, some shell scripts and little bit of perl).

With that, in 5 years, I was contacted only by 2 or 3 recruiters/contacts that emailed me because of my projects. For two of them, it was more a general look at which languages I used in my projects ("hey, this guy knows Python and Bash, he may be good for an SRE position").

The last one actually came across one of my project which is a little bit domain specific (an RFC 3161 timestamp server), not sure if it would have ended in a recruitment offer however.

I get more offers through my crappy linkedin profile.

> instead you get dance-like-a-monkey "coding tests" (unpaid, of course) that are nonportable from employer to employer.

This - the non-portable aspect - is what irks me the most.

Especially for enterprise developers. Since most of the software we develop is under NDA, we cannot put any of it online to showcase our skills.

I've had to do 3 different take-home "coding exercises" over the last 10 days. 1 was for a Hedge Fund, another for a Real Estate company and a 3rd for an Accounting SaaS company.

I'm in the process of publishing these coding exercise solutions after polishing it up a bit more onto my github profile. Still, I feel like this will also be a waste, because like OP says, most companies have their own specific "Coding tests" and wouldn't take an already coded solution to accept skills and experience of the candidate.

If you asked a musicians to perform an audition before you hired them, would you accept the response, "Sorry, I don't work for free."

Would you expect a musician to practice their instrument in their free time or only play their instrument during paid gigs?

The music industry is exploitative precisely because it has this dynamic. I would like the developers lifestyle to be more like that of surgeons than like that of musicians.

That is a bad analogy. Supply/demand ratio of musicians is way higher then for developers.

That's not the only reason it's a terrible analogy; Musicians are artists, developers are tradespeople. If an interview method doesn't make sense for a carpenter or an electrician, it probably doesn't make sense for a developer. That being said, I don't think there are any good proxies for whether a hire will be a good fit, so I understand the problem employers are facing.

I agree that it is a bad analogy, but I think it goes far beyond supply and demand differences: Music is performed mostly by muscle memory, "system 1", while code is developed mostly by ratiocination, "system 2".

Supply of wannabe developers is quite high.

Isn't that the case in all skilled professions that pay well?

I don't think so. Development seems easy just like singing you can do it at home, have some results on your own computer. But look at talent shows, there are people who can sing and win but still do not have what it takes to build career after winning talent show. Secondly there is whole bunch of occupations that pay well like dentist but how do you even practice that without buying expensive tools.

This is becoming a real problem IMHO. Asking me to do 3 to 4 hours of work on a "test" just to get to a real interview is a big ask. I do it because I have to; but as my time has gotten more valuable I've started just politely refusing and thanking the recruiter for reaching out. I have a ton of stuff on github that I'd much rather talk about.

We generally offer candidates who made it through the 15 minute phone screen to bring along some code to an interview. Quite honestly, it can be their own project or we can supply a simple "test" so that they have something to bring. We make it clear that we're not looking for perfect code samples, or even great ones, just something for the candidate to talk about. This seems to work pretty well, makes sure they're not doing any free work for us, and doesn't ask for much in the way of their free time.

I've had too many "tests" that were a giant waste of my time to ask it of others.

That's because they want to gauge you against other candidates and the only way to do that fairly is with a standardized test.

I see the point, but some of these tests can take several hours or longer to complete. I think they shoot themselvess in the foot a bit because many of the really good devs won't waste their time with that. My evidence is all purely anecdotal, however.

If all you've got is a foot-gun, everything looks like a foot.

You must be applying to the wrong companies then. I just got done doing interviews, and I looked at what the second round candidates had available on GitHub and their personal site, and the one who got the job had about 5 projects that were impressive.

Good employers care. I care. Its not signaling. Its important.

>Good employers care. I care. Its not signaling. Its important.

It's not important. There are plenty of good companies who don't do these things that manage to pay developers $100k salaries. They may assess in other ways, but it's not the only signal.

It's only important to you. You choose what you want to see in candidates. You need buy-in from the candidates that your process is conducive to a good environment, good code, and good delivery. In turn, you cultivate a team who share something in common, and the team gets along well and makes good code.

It's people finding like-minded people -- the same social movements that are in every other profession. There's nothing wrong with it, but there's currently no prevailing right or wrong way to do it, otherwise that would be the technical interview.

Practically, development is still in demand. So if I get rejected, I can go apply some place else. That's why it's not important.

Maybe your pay is slightly above those other companies, but is a marginal pay increase really worth all of the extra effort? That's up to the developer. But the employers that don't use this criteria are not non-good employers.

For me, a marginal pay increase is only a differentiator when comparing two otherwise equally ranked opportunities. I once left a 300k job to take a 180k job on the basis of work content, cultural fit, and optionality. But, I make most of my money by holding XMR, so the difference in pay doesn't determine my quality of life - nor my ability to exercise charity - which is, practically, the same thing.

Best company I worked at looked at my personal site and Github, albeit briefly, ahead of time and didn't ask much if anything about it (IIRC). They asked more peer-level questions. Pretty casual. As my first job out of college they didn't expect much from me but later told me the other devs were impressed. Best group of devs I've ever worked with. All of them were pretty smart, wrote nice code, comments, knew how to use version control, and configuration management. They wrote their own JS framework that interfaced with a Java backend. I would later realize (in retrospect) this code was really well written and maintained.

Worst company I worked (work) for (but pays well) had me answer a few questions on a whiteboard. One of which was 'not even wrong' (the answer was: "HTML5 gave us the <strong> tag to replace <b>").

They were desperate for bodies after a bunch of people left. They also partake in the usual H1B/contractor games and wonder why they have shit code. Meh, my project is going well. If they have me work on their main .NET monstrosity though, I'll quit. Haha, jokes on them.

> the answer was: "HTML5 gave us the <strong> tag to replace <b>"

i bet the question was: "what has HTML5 done for us?"


Yes, it was. And then she followed up more specifically with, "what we can use instead of the <b> tag?"

I said, "CSS, but it doesn't have much to do with HTML5's features."

It's not important. You're skipping out on a large pool of fantastic candidates who have families and hobbies outside of work. People who save their best programming effort for their job. Relying on GitHub / side projects as a barrier to entry is ridiculous to me.

But why is it important to you? Do you think the other candidates are couldn't, or didn't, produce similarly impressive code?

I noticed this as well.

The interviewer has checked the "codes in his spare time" checkbox" and I've guilt tripped them into skipping the dance like a monkey test once or twice by pointing them at my github but that's about it.

I'd be super impressed if a prospective employer looked and made some sort of salient comment, but it's never happened and probably won't.

As much as I hate it, the problem with taking the github profile as evidence of proficiency is that it's easy to plagiarize, and the applicant has a significant financial incentive to do so. Obviously plagiarizing code samples would be a terrible career strategy, but that doesn't protect you as an employer if you get the one in a thousand con artist who can collect six months' salary before people start to ask difficult questions.

Another hiring manager here who reads GitHub profiles and loves to have a candidate walk me through what challenges they experienced, design decisions they had to make, etc.

One question I have for the grandparent post: what would a meaningful question be? Is the description I mention above seem like a meaningful way to discuss the work you took the time to share with me?

FWIW, I'll look at a candidate's Github account whenever they list it on their resume. That doesn't mean I'll bring it up during the interview unless there was something particularly interesting to talk about.

It's pretty useful for screening candidates, especially looking at code written <1 year ago. I do keep in mind that code written in free time might not have the same standards as production code, but often there are some red flags that carry over.

You realize that github dates can easily be falsified, right ?

Besides, you are only using this as a negative ? So you'll actually hire people who don't put in extra effort over people who put slightly substandard free time code online ? That seems like a losing strategy.

I've always found that willingness to put in effort vastly outperforms ability. Not to an infinite extent, but quite a bit.

> You realize that github dates can easily be falsified, right ?

If someone knows that much git, I'd hire them regardless. I'm so sick of untangling other peoples failed merges or trying to explain what the "not a fast-forward" issue means when their push fails.

To be fair to the people whose merges failed, I believe this is a shortcoming of git (the tool) and not the people using it.

It is fair to expect a developer to invest sometime in learning their tools well but git is a different beast. I know it solves a very complex problem but the fact that it exposes a complex interface to the user to do so seems like a shortcoming of the tool to me.

Heck, we even have a comic for it that shows how sometimes just saving my work, deleting my clone, re-cloning the repo and applying the changes in my work is easier than understanding the Git error messages: https://xkcd.com/1597/.

The tooltip in the comic is also sound advice: If that doesn't fix it, git.txt contains the phone number of a friend of mine who understands git. Just wait through a few minutes of 'It's really pretty simple, just think of branches as...' and eventually you'll learn the commands that will fix everything.

Haha nooooo, it's really not that difficult! I really don't think git is a different beast. It's much, much harder to master a programming language or web framework.

Compared to something like Rails, iOS development, or React + Redux + Redux Saga, Git is just a couple of commands and basic concepts. Everything is listed here, and I don't think that's an unreasonable number of commands to know: https://git-scm.com/docs

The number of commands you show may not be unreasonable. But after spending many years in the industry, I think it is far easier for me to learn C++ with all its different types of pointers than it is to learn what Git is trying to tell me when I have screwed up something.

Either Git has a complex user interface like I claim or I am beginning to grow old.

> To be fair to the people whose merges failed, I believe this is a shortcoming of git (the tool) and not the people using it.

I disagree. Like everyone, I've experienced a failed merge that turns into a diff nightmare. However, when it has happened I've figured out how to solve it myself instead of wasting someone else's time with it, and it hasn't happened in years now leading me to believe it is directly correlated with git skill-level.

I think many developers just need to take more time and learn git properly. This is a good reference: https://git-scm.com/book/en/v2

Same for regular expressions. Once you know the syntax, they're pretty easy to understand. It just takes practice.

>You realize that github dates can easily be falsified, right ?

Yea, but why would someone bother in this case?

>Besides, you are only using this as a negative ? So you'll actually hire people who don't put in extra effort over people who put slightly substandard free time code online ? That seems like a losing strategy.

That's not what I said at all. It's potentially a positive or a negative, just like anything else on a resume. I've seen plenty of Github profiles that have files with 2000 lines of spaghetti code and no structure or testability. Just like I've seen profiles with neat projects and excellent code.

>I've always found that willingness to put in effort vastly outperforms ability. Not to an infinite extent, but quite a bit.

I agree with this in general too.

Anecdotally, as someone who has been involved in hiring and (sort of) cares about side projects, it's not actually the technical implementation I care about, that's true. It's about the applicant signalling "I care about this stuff". But it's far from the only indicator and shouldn't be given too much weight. I'd gladly hire someone who had no side projects but passed our technical test/interview and soft skills interview.

As for why I don't take Github profiles too seriously, we had plenty of applicants who did have projects on their Github profiles who tripped all over themselves on our simple coding test. Maybe we would have found out something similar by digging into their commit history, but that would be an even greater time investment than the hiring process already was.

I think that coding tests and challenges apply a sort of artificial pressure and requirement to have things memorized that aren't there in the real world. I think it would be a mistake to discount a candidate who has good code on their github when they had time to sit, think through problems without being under a magnifying glass, but trips up during a technical interview challenge/puzzle/whiteboard or whatever. Maybe in those candidates you could actually walk through their projects with them and try to determine if they have a deep understanding of what they wrote. To me, that seems a better indicator of how they would perform on a team with real problems, rather than having to puke up a pseudocode merge-sort from memory.

Code is for tech reviewers (CTO or inside developers). Most often, they will look at your repos, but they are not the ones you will chat with.

It's still useful to have material for non tech people, though. That's why, additionally to my github profile, I maintain an angelList profile on which I document the why's and how's of my projects from a non developer perspective.

But of course, the single most important thing you can do in an interview with a non tech guy is to show them you understand their business, what they're trying to achieve, and that you find it exciting (provided it's true, obviously). From having been on the hiring side, I've learned that the worst thing to say is : "yeah, actually it doesn't really matter what we will do, I'm mostly interested in coding".

There was one, (literally: ONE) time I check out somebody's GitHub profile when making a hiring decision. It was a clusterfuck of forks, and source repos were a total mess, no documentation, nothing. In such case I would rather NOT see a GitHub link like that, really. He didn't get a job.

It would be nice if GitHub offered a grouping function for "random" forks. Sometimes you just fork something out of interest or to fix some small bug and they mess up your profile...

I don't share your experience and my guess is that you're not advertising your projects enough.

If you have projects that you're proud of — mention them on your website, on your Twitter's bio, on your LinkedIn profile, mention them everywhere. Even better, once per year go to a conference and talk about those projects in front of an audience, such that you can then have videos on YouTube too.

We like building things and there's absolutely no shame in marketing ourselves or in talking about the things we've built.

I think maybe there's a miscommunication here. Of course I've had conversations about my projects. Of course I've discussed them at length and in public spaces. Prospective employers have never asked. And, yes, it's clearly highlighted on my resume and my LinkedIn profile, with an entire section of my resume calling out interesting and high-quality projects.

Employers do not care. Other people? Sure. Employers? No.

There are also employers that downplay your achievements just to have a reason to offer you less money. The decision to hire you was already made and they advanced to salary negotiation but without telling you.

You just have to know your real market value, how good are you compared to others.

Well, yeah. That's why, now, I run my own little consultancy. ;)

Thanks for mentioning this. I just finished a desktop app. They wanted to pay X but when the little app was done another colaborating company found out so I made 2.5 * X in total. Now the first company came back and they want another app for 4.5 * X. Is this how it all starts? :)

Pretty much. Making the coefficients go up to the point where you can hire other people is the name of the game.


I have put my GitHub + GitHub projects on my resume as individual items with clickable links, but very few employers (mostly nontechnical recruiters) acknowledged them.

Interesting. Hasn't been my experience. I've hired and been hired based largely on github code many times.

I imagine it varies. I have gotten some good mileage out of my github projects in interviews that have resulted from the HN whoishiring thread; I'm sure if I went in for an interview at some line-of-business software shop here in Houston this would only result in a blank stare.

So when I was interviewing as a intern and junior developer, I had _most_ places that actually interviewed me look at my side projects I'd listed on my CV. The one interview I've done at an intermediate level (at a company that reached out to me) also did.

I agree with this, the whole need for a portfolio for a back-office role is bizarre to me - no one who has asked for my github has asked me about them, they are much more interested in me talking through my work related projects.

I've been taken aback during interviews and client conversations when they brought up details about projects I'd posted on my off time. There are people who look at that stuff.

Just to add my data point, my company does look at side projects as part of the hiring process.

That's too bad. This is my favorite interview question to ask.

same here, have github full of projects. Nobody ever mentions them.

One company I work for rejected a candidate based on Github - he'd cloned some repos and stamped his name on it, and claimed they were his. The hiring tech smelled something fishy, and the candidate said that he worked with these various other repos on the backend. Hiring tech reached out to the owners of several repos, and none of them had heard of the candidate before.

Github was used in this hiring process to reveal a fraud - a different 'success' story to others commenting here, but it definitely birthed 'a meaningful question' or two.

Likewise I've deleted resumes because they claimed to have coded a website when it's clear they simply installed out of the box software. They didn't explicitly lie and say they coded it from scratch, but I felt like they were implying that

Another time I rejected due to side work was a candidate claimed to have written the "worlds fastest framework". Even if it was fast it's possible there's something out there faster, so it showed lack of touch with reality, ego, and lack of self awareness. Note I gave the candidate the chance to substantiate their claims and was met with profanity

> so it showed lack of touch with reality, ego, and lack of self awareness.

feel like they'll fit right in with any number of startups out there! But good for you on calling them on it.

I am technical and we will sit down and go line by line, haha. And I know where alllllllll the fun code is hiding, not just the boilerplate. I get real deep in there. I’m sure my methods probably weed out some folks who don’t or can’t drop trou, but I never expected to catch every single fish in the ocean.

You're espousing a false dichotomy. The options in evaluating a candidate aren't merely "give them whiteboard coding problems" or "examine their side projects". Give them a take-home assignment. Give them an in-person assignment where they have several hours to complete a smallish design-and-build task. There are a lot more options, and I'm getting tired of people claiming they don't know how to evaluate technical candidates when it seems like people are just saying "whiteboarding is a bad test" and throwing up their hands and giving up.

Sure, you can use side projects to complement other data, but if you're relying on it as the primary measure of a candidate's programming skill, you're doing it wrong.

Giving a candidate a "take-home assignment" is actually the most disrespectful of the options listed.

Unless you're an awesome company, with really cool projects and offering me a jaw dropping salary and benefits, then no, I'm not going to do free work for you. Even if you are an awesome company, people probably have to actually work for you to see that, so no, I'm not going to do free work for you.

I'd rather spend those hours doing work that I expose on GitHub.

> in-person assignment where they have several hours to complete a smallish design-and-build task

That's another proxy that's as useful (or as useless) as a coding exercise.

I don't know what kind of homework assignments those companies give but I can't even imagine how something a candidate could code in a few hours could have internal use. It takes days for even new hires to start contributing useful projects, let alone a never-before seen candidate without knowledge or access to the codebase and internal systems.

The homeworks are mainly a time saving device on one hand, where they replace an hour of phone screen time for an engineer with hopefully less than 30 minutes of grading. For the candidate, they may take a couple of hours instead of 1, but they also allow for a more relaxed setting with use of IDEs, Google, etc. and no interviewer watching the typing and tapping their foot.

"Free work" in this case doesn't necessarily mean "free work which is useful to the company" it just means doing similar things to what I'm payed to do normally for free with no benefit to myself.

I've gotten complaints from companies where I've done homework assignments for interviews that it didn't seem like I put enough effort in and my reply has always been that I get payed good money to do this stuff at work, they're getting the left over hours in the evening when I'm of course not performing at my peak. If they want work-level quality they should pay me for the assignment.

That's fine, I'm happy to hire someone who is dedicated, passionate, and loves coding enough to to a quick assignment. Its not free work, its a demonstration of skill.

In fact, we just hired someone who did that, and her assignment came back better than expected. At first we were so-so on her due to potential lack of experience, but the assignment (and her projects on her personal site) showed us were wrong.

>Its not free work, its a demonstration of skill.

Was it paid? If not, then it's free work. Call it whatever you want, in my experience the people talking about passion the most are those that are just looking to pay the least they possibly can.

Why not call it an audition?

Until we have established a name (track record) for ourselves, and/or have a number of respected people vouch for us, we are called on to demonstrate our skills, just as musicians and actors are.

In principle, I think this is reasonable, especially since there are reportedly many fakers and exaggerators applying for jobs.

In practice, it appears that there are so many shoddy auditions that we must endure when looking for a job, that it becomes almost a full-time waste of time. This is unreasonable, methinks.

The problem will become serious when nearly every company that is hiring runs a low-quality dog and pony show in which we must perform, or not be hired. We can already see an analogous issue in the ubiquitous presence of objectionable binding mandatory arbitration clauses in seemingly every contract we sign nowadays.

If you were hiring a musician, and you asked them to perform an audition. Would you accept the response, "Sorry, I don't work for free."

"Well do you have any recordings you can show me instead?" "No, I do other things with my free time. I don't work outside of work."

Auditions don't usually take several hours, do they?

That's another reason the "audition"/side project analogy doesn't work. I'm a musician. When you audition for a band, it's a matter of walking in, plugging in or setting up, and playing a few songs with the band.

If the musician is any good compared to the people they've already auditioned, many times you won't play another song. The band will hire you on the spot.

When you hire someone to fix your roof, do you make them build you a doghouse first to prove they know carpentry?

If I owned a carpentry company, and I was trying to hire a carpenter, yes, I would ask them to build me something. Maybe not a dog house, but a bird house. You don't think that would be reasonable? "If you want us to hire you as a carpenter, prove to us you can build a bird house first." That seems totally reasonable to me.

If I hire them to fix my roof and they leave a huge hole in it, I am expected to sue them and wait for a few years to recoup my losses (while having a huge hole in the roof).

I think I prefer the coding test.

So you are admitting you are bad about making decisions about people? Luckily you have something empirical because you are no good at judging someone. Thanks for admitting that, the first step is accepting you are no good at evaluating people.

Ha, but seriously, I hate take home assignments.

> So you are admitting you are bad about making decisions about people? Luckily you have something empirical because you are no good at judging someone.

This is almost entirely the point of take home assignments. In fact, this is almost word for word what my response would be to anyone who doesn't like take home assignments!

> Thanks for admitting that, the first step is accepting you are no good at evaluating people.

Step two is admitting that, in general, evaluating people using non-empirical means is an enterprise fraught with peril.

I largely agree, but I have had one experience that opens a tiny bit of wiggle room for such assignments. During a phone interview with hiring manager, he asked me if I had experience with X, Y, or Z open source projects (all relevant to the team). I told him that I did not, but that I'm always eager to learn new things. He asked me to pick one of them and 'build something' that uses it (no more specific than that). I instantly recoiled at the idea, but I didn't verbalize it. I went along with it. I decided to take the 'when given a lemon...' point of view and build something that was of interest to me.

A few weeks later I came onsite for in-person interview. I brought up the assignment and was able to use it to show my interest and capability of quickly learning and putting something to use. I really believe it worked to my advantage.

If the assignment were to implement something very specific I would not have done it. In my case the assignment was so completely open-ended and was able to convince myself that it wasn't a 'work for free' situation.

You can actually achieve a happy medium between the two. About 2 years ago I interviewed with one of the large SF-based enterprise companies. They emailed me a cool little assignment (parse a text file that describes dependencies and build a graph, output install manifest) and asked me to send back a solution in an hour and half. An engineer was available to call on the phone to answer any questions, but otherwise I was left alone to code in peace, with no screen sharing or other kinds of watching as I coded.

It's probably the best phone screen/coding assignment experience I've ever had. No pressure from someone looking over your shoulder, but limited in time and scope, so I didn't spend all day working on it.

> I'm not going to do free work for you.

It's kinda funny that so many people have this attitude, when, as a salaried employee, we likely have all done more than a standard 40-hour work week on occasion (or habitually), without extra compensation.

Do you also consider time spent prepping for an interview (brushing up on algorithms, learning about the company, researching interview questions, whatever) as "doing free work"?

I just really don't get this attitude. It's not "work", it's a evaluation. And it's an evaluation that actually displays skills relevant to the job, unlike many other approaches.

Most people already do not like brushing up for algorithms. Learning abou the company should take 20 minutes at most. When these assignments take hours of your time with little to no feedback (from my experience), it feels like busy work.

Depends who you're hiring. Early career devs with a lot more time than prospects, and who have already spend a good portion of that time working for free on exercises and toy projects in the name of learning, might not have a folio of work to prove their skills. Crusty HN regulars probably prefer to chuck a github profile around, or provide samples of work in some other way. I don't think it's one size fits all.

> I'm not going to do free work for you

Our industry could follow after the restaurant industry and implement stages/trialing. You come in and work a full day. If you get the job that days pay shows up on your first paycheck, if not they send you out the door with a check.

That does not work for anyone who currently has a job, unless your current employer has extremely lax moonlighting policies.

True, but a paid short-term contract to do an actual task for the company would work. Knock out a small feature during the night/weekend. Keep it something contained and relatively straight-forward. If they can get it done, it demonstrates the candidate's ability to parse your codebase (so they are learning something they'd need for the job anyway) and they have contributed.

I think most companies just don't want to take the time to set this kind of thing up. Easier to just pass on people that won't do the requisite song and dance.

No it still wouldn't work for someone with a current job. That's a huge risk if you're responsible for supporting a family. If you quit and the new job doesn't work out, then you're kinda hosed.

Its a good idea but speaking as a non-citizen on a work Visa such kind of work would technically be illegal for me to do.

Doesn't really work for any larger scale product, it's unlikely that any candidate will be productive in their first day. Sure they could build something separate but then the value for the company probably diminishes. Unlike serving dishes the process of making software differs quite a lot between workplaces.

That's not entirely true. The way kitchens are set up/configured, what goes into dishes, how stations are set up, etc can change quite a bit from place to place. Additionally there's usually kitchen slang/words for most items on the menu which need to learned.

Also I've had many dev interns at my company, it takes work on my end to enable them to be productive quickly, but it's not impossible. Maybe you need to write some skeleton code for them to fill out and unit test, maybe you need to document things well and spilt up your tasks well, but you absolutely can have some one come in and be productive day one.

In fact my interview was a college hire group interview, we were given a laptop with skeleton code in multiple languages, a problem to solve with fixed inputs. The problems were similar to what you'd likely encounter for whiteboard coding, except I got to be left alone with a computer, the internet, and my editor of choice to write actual code to solve the problem.

> so no, I'm not going to do free work for you.

My current employer, when hiring me, said "Have you ever looked at the Firefox codebase? No? Good. Here, take this Firefox bug. Fix it by Monday, or at least analyze the root cause, and let's talk about it during the interview".

Needless to say, they had no interest in Firefox per se. It was just a cool open-source project, and a good way to give me "actual work" while making it plainly obvious that it wasn't for their own benefit/ they weren't looking for "free code via tech interviews"

This seems like a weird position to take from the employee's side. I'd much rather work a few hours on my own schedule and on projects I choose and that I can then show to multiple companies, than to have to write multiple custom hours-long projects for each company on a fixed day.

I would say take-home assignments complimented with an assurance of an in-person interview is the trade-off that I'm willing to make. I recently interviewed with a company that gave me one and told me that my solution would be the basis for the final onsite interview, which made a lot of sense to me.

After I submitted my solution though, the company came back to me and said that they had received a lot of submissions and would not move forward with my candidacy. This made me intensely angry as I had spent an entire evening coding up the solution.

To add insult to injury, they even asked me to remove the solution from Github where I had posted it (I had not posted the question so it wouldn't really make sense to someone just glancing at it). At that point I stopped all communication with said company.

Yeah, it's not ethical to give those out without basically already having committed to interviewing the candidate.

Likewise the companies that think they are more important than my private life are not worthwhile for me to apply.

If the HR isn't able to evaluate my skills from my references and the projects I helped my employers and customers to earn, there are also lots of companies to apply for.

Just like employees, companies aren't special snowflakes.

Besides most of my github stuff is useless for my type of work, because I am forbidden by contracts to publish anything related to our projects.

I don't understand the haughty attitude with which many companies treat potential candidates. And there are some truly terrible ones out there. Its worked well for me since I wouldn't want to work in a place that treats potential future employees that way. But in most big orgs, HR and hiring is somewhat divorced from the actual day to day so I keep that in mind as well.

> unmaintainable junk

Isn't that what pretty much every side project is anyway? Unless you're an active OSS developer the chances of any of your projects being worthy of judgment is pretty slim.

I've got 10-12 repos on github I can show you...they're all half baked weekend projects that I did just to tinker with some new technology or idea I had. None of them are production quality, and most of them I'm appalled at anytime I go back and look.

none of my private doodles have gone through design review, actual testing, user feedback, deployment, and the 500 (or 1000 or 5000) closed bugs it takes to make them representative of professional work.

neither does that 1 wk coding exercise you made me do unpaid which I'm pretty resentful of.

If you have an open source project and want me to fix a couple bugs and do some reviews to see what kind of work I do, thats pretty fair (Rethink asked me to do this, I couldn't afford the time, but I thought it was a good plan)

Otherwise, I really don't know what you expect to learn

A lot of programming is real grimy and there’s a lot of skeletons in every closet. Par for the course. I’ve seen it all. The code isn’t the code, it’s a conversation we have about how you think. So many people dragging me in this thread think it’s about the code. It is not. I’m interested in the person. The code is just an artifact and a conversation piece.

Then certainly it’s replaceable, no? Could just as easily say it’s not about the whiteboard, the whiteboard is an artifact and a conversation piece.

Most assuredly. But what are we talking about on a whiteboard? Not the things that I personally am interested in, or the things in my opinion most important to the success of a project.

Honestly the code review isn’t for the obvious rock stars. Anybody can pick them out. It’s for the sleepers who might not get a fair shake but who can deliver.

I feel the same whenever I have this discussion. The "Look man, I have a life and I don't want to do extra unpaid work just to get a job" attitude is misunderstanding the motivations (to be fair, there are some situations where they are not).

I look at what people do outside of work that isn't writing code. Having something means you are probably healthier, less likely to be burned out and more likely to have experiences that let you better understand what solves real customer/client/user problems. I don't want to hire someone who only writes code, they aren't as good at solving problems.

Now I have to know if you can actually code. The best way to do that is to look at some code you've written. That's basically the extent of it. I'll do some extra work to come up with a take home project if I have to but if there is a great candidate already, and they have a github with good code in it ... I have to manage my time and I'm often chasing them when hiring, not the other way around.

Open sourced code on Github is just the most common way to do that, and it does get bonus points for having examples of dealing with actual users in issues and involves giving back to the open source community that you almost certainly take from.

But honestly, I really just want to see some actual working useful code you wrote.

> Having something means you are probably healthier, less likely to be burned out and more likely to have experiences that let you better understand what solves real customer/client/user problems.

I didn't learn to write code until my late twenties. I have a fine arts degree. My early and mid twenties were focused on activities unrelated to technology. I have a reasonably large amount of life experience unrelated to programming. I haven't found that this experience gives me an edge in solving problems in my job as a programmer. Can you provide concrete examples that demonstrate this idea that people who code less outside of work are more effective problem solvers? I've seen this truism floating around quite a bit lately but have yet to see it demonstrated.

Not that you code less outside of work, but that you do other things and meet other (non-programmer) people.

As to how that helps: it doesn't help you solve implementation problems, it helps in designing the software to do what the client wants in a way they will like.

It is hard to put yourself in the shoes of a different person to design something for them if you only hang out with people like you and only use software built for people like you.

As a concrete example: what would a piece of software for artists be like if your only-engineering coworkers designed it vs. you having some input? By the way, it would definitely be either a portfolio hosting service or some kind of art database software. What problem do artists have that could use some software help that they aren't even aware exists?

The reason you have a hard time with it is because often the skills acquired through disciplines in the arts are less tangible then something technical.

If nothing else, you might be more interesting to work with, have a clearer communication style, better taste in evaluating what "looks good", etc. etc.

These aren't a substitute for skill, but past a certain skill threshold other things start play more significant roles.

Edit: I really don't understand this pattern of down voting comments without offering a rebuttal of any kind. What exactly about this comment was against HN guidelines? Last time I checked, disagreeing with a comment wasn't grounds for down voting.

I agree and would like to emphasize all of your points. It's somewhat ridiculous not to hire someone because she has no line of code to show, as there are plenty other ways to poke at their knowledge.

But I'm bewildered that this person has zero line of code to show. The only case I could think of where that makes sense to me is that the developer is still in college. Imo even recent grads generally have something to show off for what it's worth.

You don't even need to be hacking every night and week-end for that to happen. If you're a minimum into what you do, chances are you put at least a night or week-end aside to hack together something.

Personally, I'm far from behind a great developer, but you can find loads of "unmaintainable junk" in my GitHub (github.com/hery). And as many pointed out, when you're prolific, you tend to accidentally produce valuable things, and I do have some decent projects in there that I explicitly pointed out in the process of interviews, which helped me skip technical challenges or interviews straight to offers. A bit extreme huh?

Most developers have a life and hobbies outside of coding, and I believe it holds even truer for the best developers. But with decent priorities, it's easy to invest some time in hacking side-projects. Otherwise it feels like you're putting 99% into other hobbies, and 1% into coding, i.e. you're just doing it to make money. Balance is key.

Eh, the Duds can fork github projects too, add some minor feature and release the "project" to rot.

You do not get to see people who get stuff done- you get to see people who optimized their "work"-life for the impression of getting stuff done.

Not beeing ready to pour unreasonable hours in- actually is a good sign. Such a person values his/her time- and thus will think on there own, for example when handed ridiculous feature requests. Such a person will stand up to me, and ask questions that might save month of development time sunk into unnecessary work.

Sorry, but we are not building the pyramids here. We are making software- and for that a lazy-loading brain with a clever work-avoidance Algorithm, in a un-passionate mercenary can be very effective as a development tool.

> the Duds can fork github projects too

Sounds like a boogeyman.

Since you can see right there in Github's UI that something is a fork and that they are not the main contributor to the project, the dishonest applicant would have to go through some trouble to fake it as a sideproject.

The same dishonest person can also just lie about their work history.

Eh, abandoned forks will just have all "6 years ago" under the updated column. It's not a good look as I see it, but it's not hidden.

I think the author is confusing the search for passion with the search for better information. If you can’t share code from past and current employers, and you don’t believe that your quick and dirty solution during the interview represents your best work, then a side project that shows a little carefully crafted code is one way to give a potential employer better information. And better information derisks hiring decisions. With perfect information, hiring decisiInd make themselves. It’s perfectly fine to prefer art projects to side projects in code, as long as you also accept the informational disadvantage and its repercussions.

Everyone is responsible for showcasing their skills and selling themselves. People who refuse to do that are only hurting themselves.

There seems to be a weird inversion in the minds of many commenters here: that they are somehow entitled to dictate the terms under which they are selected to work somewhere.

Isn't the point that if you want the job, you do whatever it takes?

Why are we not celebrating the fact that we work in a kind of meritocracy where it really is possible to showcase your skills without any credentials or reference to your past?

All of this bloviating about "wasted time" in the context of demonstrating skills reeks of bitterness from lost opportunities to me.

People who refuse to do that are only hurting themselves.

If an employer passes on candidate for the wrong reason then the employer is hurt as well. A good hiring process shouldn't rule out people for reasons that aren't related to their capability to do what they'll need to do in the job, and very few jobs require a good github profile. It's a useful proxy for development ability and passion, but that's all. You can be an awesome developer and never touch Github.

If you want people to show up with code why aren't you asking them to take home something and show you their solution in the interview, working through it step by step?

I guess I see that as disrespectful of their time, but I could probably work something like that out. I run a small shop. I am allowed to use my good judgment.

What about the time it takes for interviews? Is it being disrespectful to take up an hour of their time for an interview? What if you do a 2nd interview for another hour? What if you have a 3rd? It's not unusual to have 2 or 3 interviews before hiring. What if you replaced one of those hours with a coding task that should take under 1 hour to complete. Is asking for 1 hour disrespectful to their time? Why is asking for 1 hour of their time to code different than 1 hour of time for interview?

I think the expectation that I have random projects that I'm passionate about to be more disrespectful of my time. I have to compete against an unknown, I don't know what the other people interviewing will be doing for their projects so I have to try and make it more impressive than projects I don't even know about.

It's much less stressful and takes much less time to give me a basic to intermediate coding problem to solve on my own with a pre-set number of days/hours to work on it.

I think many of the people on HN that say they won't do a take-home assignment would actually do a 30-60 minute take-home assignment if it meant avoiding the "code some algorithm on a whiteboard while the interviewer glares at you" situation. Then on the on-site, you can focus on the design/behavioral type things appropriate to the role's level of seniority.

Also, if you give the same assignment to multiple candidates, you'll get a better sense of how people use different approaches to solving a problem. You can still use discussions around side-projects to complement the process, but having a standard approach for all candidates is helpful.

> I think many of the people on HN that say they won't do a take-home assignment would actually do a 30-60 minute take-home assignment if it meant avoiding the "code some algorithm on a whiteboard while the interviewer glares at you" situation.

Except it never means this. It's always "and" and not "instead of". Many companies do these irritating programming challenges now, and then when you pass them, will still bring you in for the standard Whiteboard Hazing sessions.

I needed to emphasize the if in that statement. :) The typical saltiness regarding interviewing on HN is warranted since it's usually an "and" thing.

Most people would not think twice if a potential employer interviewed them for 30-60 minutes and then asked them to do a follow-up interview. It's not that unusual to have 3 interviews before getting hired.

But if you ask them to do a 30-minute coding task, that suddenly becomes a total disrespect of their time.

>I’m not mad at you, but as the dude on the other side of the desk, I have to decide whether you can cut it and I’m fine saying no. I could make you do some whiteboard problems, but I think whiteboard problems are pretty far removed from your day to day development and I am uncomfortable relying on them as a proxy for your ability to ship code.

So, are architects and engineers expected to have built side-houses, and side-bridges, and side-cars?

Architects and engineers can usually show you what they've done in the past. Software developers can't if they don't have side projects.

Artists have portfolios. Why shouldn't software engineers?

Because artists are usually sitting on their asses waiting for work, not already working 60 or 80 hours per week churning code.

If you're working 60 or 80 hours per week churning code, surely you have some code to show. </sarcasm>

Really now, the fact that you can't showcase what you've done for your previous company is an implementation detail. Using a portfolio to judge a candidate is perfectly valid and works great with other industries. The reasons why you don't have one are your own concern.

Like I said in other comments, interviewing is (for better or worse) a game / competition. You need to optimize for success.

>If you're working 60 or 80 hours per week churning code, surely you have some code to show. </sarcasm>

Yeah, if you work for GNU or for countries were NDA is not a notion </sarcasm>.

Dude as a Dude who knows a little about employment law all your work code belongs to your employer.

Yep. I've built quite a few useful shell scripts over the years for various employer tasks or demands. Only one of those have I taken with me, which was a quick 5-minute job to figure out what was causing a system to crash in absence of a significant monitoring solution that could give minute-by-minute specs.

At my last gig I spent a few days writing and troubleshooting a script that would perform automated tasks for setting up servers for DBAs given their specs and demands. Easily one of my most shining examples of making bash work for me. It's back with the employer since it's their property.

That's actually not always true. Usually, yes. Not always.

what you do at work absolutely is - there's a question about non related work done off the clock and with your own equipment - but its unlikely that most have the $$ to be be able to fight a lawsuit

Sure, sure. And if you followed the letter of the law I’m sure you deleted it all from your laptop promptly and is unavailable to you no matter what.

But if you didn’t, for whatever reason, and you happen to have an illegal copy of some production CRUD app hanging around so I can look at how you organize your thoughts, nobody I’ve interviewed has ever gotten even a whiff of heat for letting me look over their shoulder and see how they wrote their schema. I don’t want a digital copy and I could give a shit about your illegally retained IP (and there’s a whole lot of it floating around). I want to know what you name your variables.

The best way forward is absolutely some open source code, folks clean up better in public and you don’t have to make apologies for things that weren’t your fault.

But this smells like folks just wanting me to take their word, and it ain’t gonna happen. I’m flexible, but when in doubt I choose ‘no’. If the next guy can prove it and you can’t, hey, plenty of fish in the sea.

A candidate showing me code they wrote for their previous employer would be pretty much an automatic black flag for me.

It's a matter of "if you want to know how they'll treat you, look at how they're treating their ex".

exactly. I work damn hard when I leave a place to make sure I have absolutely zero IP. its likely none of it would be useful to anyone anyway. But thats what I agreed to explicitly, thats what my former employer wanted, and as unlikely as it is that anything would happen I wouldn't want to taint a new company with code that can't be clearly attributed to them (which is what they paid for).

thats insane

I was kind of with you up until this point. That's a line that is foolish to cross.

Any candidate who plays fast and loose with their previous employers' code is going to play fast and loose with _your_ code. Or your corporate secrets. Or lifting GPL code. Or pretty anything else they think they can make off with. Such employees are a lawsuit waiting to happen and are best avoided, as are employers who tolerate such employees.

You start off relationships with potential new subordinates by asking them to break the law in order to demonstrate their value to you? Wtf man.

I don’t know their arrangment with their old employer and I don’t ask. Very often folks will have an ongoing maintenance consulting relationship and retain some IP just in case. More often folks will just have some stuff lying around, and they’re already breaching their contract by having it.

Believe it or not, there’s next to nothing I’ve ever seen that was the remotest bit interesting on an industrial espionage level. Just regular old business logic that was either cleanly done, or not.

I honestly could not recall any of the crap from the interviews I’ve done beyond a sense of how people thought.

Anyways, you wanna do some interviews your way, be my guest. I will happily send the guys who don’t want to do show and tell to you. I’m not trying to be all things to all people. No open source, no closed source, no side projects? We can play API bingo or code a b-tree, but prob not a culture fit.

I have the feeling that the point was the simplistic and damagingly narrow screening process here not the criticism of screening as practice, quite the opposite. The side project request became so mechanistic that everyone can expect it but the reliability is questionable as the authenticity of the product is beyond validation. When people face something expected in a competitive situation then the really competitve - and unethical ones - will go to great lengths for raising their profile even through manufacturing reliable looking evidence (doing it continously, possibly plagiarizing), posing as excellent through acting. No good and easy - and cheap - screening process is available as the reliability of workforce is multidimensional beyond the human capacities to comprehend, also beyond the technical level. Over the years I had bright but unfit colleagues (damaging the team effort, jobhoppers, unsteady performance, cheaters) as well as average ones with grey personality carried the company on their backs. The real test of the bread is eating. They have to show in the particular environment if they are capable in that very specific situation or not. Side projects will tell nothing on this. Using it as a turning point is very ... to be polite, unwise. Still not questioning the need for first filters so the chance of wasted efforts is minimised by not nurturing someone inadequate into the precious environment for months in vain. But that first filter must be way broader than the topic of side projects.

Sure. But you seem to be approaching interviews as a one way street like, "I hold this thing here which I will give it to you if I deem you worthy". That is not how things work these days in tech, especially in hot markets. If you are not willing to carve out time there are plenty others who will. Sorry dude, just the reality of the rare market situation where the power is balanced between companies and talent.

I agree 100% with this viewpoint. I would much rather sit down with a potential hire and go over some scripts and projects they've written than go to the whiteboard. I don't care how big the project is, I think it's just much more pragmatic to share ideas and stories over code than algorithmic gotcha questions.

I'm sure you realize that it takes 6 months for the average person to become productive at a new job. If you can't factor that in to a new hire you're probably working at a typical coding sweatshop.

> Nor am I going to go on some weeks long hiring safari where I hire you for a week or two and see if you can cut it. No time man.

So WHAT weeks long safari do you go on for your company? Product management? Raising fund? Sales?

On boarding. Training. Professional development. Pair programming.

It’s much more cost effective to invest in your existing people. That’s one thing I never skimp on.

The best piece of advice I got about hiring was that the interview is to find a reason to say no. Too many people want (expect?) the point to be to find a reason to say yes. For better or worse, some companies will use lack of OSS contributions or side projects as a reason to say no.

If you hire explicitly based on code with no way to gauge ability to collaborate with others, design, or probe business requirements, I really wonder how strong your team is and how competent you are as a manager.

Our client always ask for code samples. Code samples we use from projects. When we hire a dev. The dev needs to at least have a github and "some" code samples. its what the client wants ( i know ignorant but here me out)We work with the project owner, who's boss ( signs the check) applies pressure on him ( blames him really) for any fuck ups in code / app failures) Meanwhile the boss changes the app features on a daily basis. So the PM does not know any better since he has to have the code approved from another teach lead. So if you want to make money you have to provide ( proof of work / code) which may take the form of a side project ( demonstrates your not in it for the money without passion) I don't know, just my experience with hiring a dev.

Do the side projects have to be domain specific? What if I were applying for a front end developer job and had some hobby projects doing GPU shaders or a toy database in C?

I work on internal tools for a major company. You can't see them, some I probably shouldn't even tell you about.

When I get home I play video games and watch tv and play with my dog. sometimes I'll screw around with a new language or something else. I don't post stuff on github and keep a blog or anything else like that.

I think your view is pretty myopic. Ask me to talk for 30 minutes about one of my projects and force me to answer some tough questions about it. whiteboard some designs.

Maybe I should just hire some guys off a freelance site to write me some side projects.

This will sound flippant, but you can’t win them all.

There are hordes of false negatives in any hiring process, anyplace. My process is more accurate IMHO but it’s indisputably more work and it does not scale at all. Ultimately everybody is going to beef with any criteria they don’t align well with and I have to remain flexible enough to not throw the baby out with the bath water. I don’t mind getting drug a little on HN— I have a good team, and their open source contributions might not be as impressive as you assume.

> When I get home I play video games and watch tv and play with my dog. sometimes I'll screw around with a new language or something else. I don't post stuff on github and keep a blog or anything else like that.

Presumably you also don't think people should stop thinking side projects and public code is important in hiring. You've made a choice to not have these things and asking the world to change one of the few sane things in hiring because you just don't fit the mold is a whole other story than just accepting that this is something you lack that could potentially sway an employer.

As another comment said: You can't win 'em all.

By chance, I happened to have a pretty productive few years on GitHub and having no degree I still managed to land the first job I went to interview for, in part because of my GitHub profile. Am I "for" the current situation because it benefited me at the time? No, because I think it's a clearer indicator that a colleague won't be a nightmare to work with than a degree or if he can implement the correct data structure on a whiteboard.

As somebody who has been on the hiring side (not the hiring-manager, just the technical person assigned with deciding whether someone looks like they are what we want or not), I've often just been given a CV after they've been 'pre-screened' by HR.

The issue is I'm usually given a bunch of CVs, and I need to fit this in to whatever other work I've got going on (probably too much already which is why we are hiring). I can arrange interviews, but that probably means I'll need to stay after hours (a lot of people can't just take an hour off in the afternoon to go for an interview), so it's in my best interests to filter down that list. Plus arranging interviews won't just happen the next day - although we need someone yesterday, just deciding who we want to hire is going to take another few weeks.

If I can see that you've got some open source projects or a blog or anything to show how you work, it really helps to make you stand out from everyone else, because most people don't have anything to show. If I had a candidate that looked good, I wouldn't throw them away because they didn't have any side-projects, but if I've got two candidates who both look good on paper, that probably would be the deciding factor.

So yes, hiring someone to build some side-projects for you might be beneficial. At the very least you could use it as a management side-project :D

Another place I worked at gave a small programming assignment (build something to import a CSV into a database). It was pretty simple, and would take you no more than two hours, but it was rather good to see how people work and whether they could follow instructions (>50% of candidates ignored the "write tests" line). This was used to screen people for an interview, during which we would ask them questions about it.

You work at a major company, you have some different quality signals available to you that can obviate all of this.

It's funny that the tech industry is so insistent on side projects. I mean - would you hire a marketing person based on their 'side projects' or a corporate lawyer purely based on their pro-bono work? Likewise, who on earth asks a building contractor if they have any side projects? You would ask for references or find a contractor via someone you know. What about recruiters themselves? How does one judge if a person is going to be good recruiter or not? What about sales guys? Does anyone ask them - do you sell stuff on the side - Y'know maybe you just do door-to-door selling as a hobby. Or what about other engineering disciplines? Does one really ask a Mechanical Engineer or a Civil Engineer about their side projects?

Side projects are for software what a portfolio is for other creative industries. You don't hire a designer sight unseen -- usually you ask for a sample. It makes perfect sense to expect the same of technology.

Only if you believe software development is neither engineering nor science, and is purely creative.

It isn't, though, and demanding a portfolio is almost as terrible a practice as Cracking the Coding Interview style interviews.

> Only if you believe software development is neither engineering nor science, and is purely creative.

I have to believe it "is neither engineering nor science, and is purely creative"? Only those extremes? I believe it is some of all 3 and I don't think requesting (i.e. not "demanding") a portfolio or prior work is any more unreasonable than not asking for it.

Fine, then you must believe it's mostly creative. Far more than you think it's engineering or science. Otherwise a portfolio would simply be at best an interesting conversation piece and not at all indicative of ability.

Why do you think a portfolio can't tell you about someones engineering abilities? Can engineering not be judged by looking at source (and potentially surrounding documents)? What would be missing for you to do that?

In traditional engineering disciplines like civil engineering, organizing bodies and licensure communicate competency, applicability, and ethical awareness. Software development has no industry-wide equivalent. There are small certifications for Cisco et al but for the general practice there's no standard.

This frequently comes up in discussions like this. Licensure is a joke with respect to demonstrating competency. Yes, the initial hurdle is high. However, once you are licensed it's hard to lose it. I would never rely on licensure as a measure of competency or currency.

On top of that, many electrical and mechanical engineers are not licensed because they don't need to be. (Everybody likes to jump to the civils in these discussions and forget about us, since civils are basically required to be licensed.). EEs and MEs don't get put through the same bullshit that this industry puts developers through.

This is true, and I regard that as a reflection of the immaturity of this industry.

I hear this a lot but I think it's deeper than that, such that letting the industry age isn't going to lead to the same end. The issue is the very one that is in this thread: programming is at a weird intersection between art, science, math, and engineering. Credentials and licensing matter a lot less than simple output. You can gain a ton of experience in the field even if you're stranded on an island with no internet and just a solar powered laptop capable of executing new code. Even without a computer you can still work like a mathematician and propose problems, even creative problems with no engineering applications, and then solve by coming up with and writing down algorithms for a computer you don't have. You can imagine arbitrary business details that get in the way of a clean program and force you to start thinking in architecture and system design.

Demanding a portfolio isn't a great hiring strategy simply because a lot of people's best work isn't out in the open. Who would demand a portfolio from someone like Mike Acton? It'd be enough to know what he's worked on and talk to him. And at the same time requiring a licensing proxy is no help either, he never went to college AFAIK and that'd probably be a minimum requirement for most actual licensing proposals.

Software development is all of those things. You need the engineering experience to know what to do. You need the science to know ways to / not to do things. And you need the creativity to put them together.

Is your "portfolio piece" spaghetti code or well-thought-out modules? Is it doing something recursively that shouldn't be? Do you have elegant code or brute-force solutions? Is it write-only?

All of these can be seen in a portfolio or side project if you choose to look.

> Only if you believe software development is neither engineering nor science, and is purely creative.

It's certainly not a science, and it's very rarely engineering. About the creativity, it's a craft. There is some creativity baked in, but about the same amount as for mechanics or carpentry.

Maybe software development should be considered a trade?

Certainly. And it should be taught like trades were for centuries: with apprenticeships.

Architecture isn't purely creative either, having overlap with engineering. You ask architects for a portfolio too.

That's more like asking them for a resume. A portfolio of side projects isn't remotely the same thing.

Software development is a craft, not fully art and not fully engineering. It definitely is not science.

I can take a 10 minute look at a designer's portfolio and get reasonably accurate insight into their style and abilities.

Do you think a 10 minute scan of an arbitrary GitHub account is the same? I say decidedly not.

This is what I don't get about this requirement. I guess you can quickly see if a developer uses design patterns, writes tests and how easy the code is to understand. I don't think that's necessarily mind blowing though and you could still get stuck with an individual who checks all those boxes and still sucks.

You are thinking about this in terms of one applicant. When you get 200 applications it's very useful to be able to quickly weed out people who are obviously not a fit.

You can also just throw out applicant number 26 and up. You most likely have a few solid candidates in the first 25 worth a phone call. And if not, look at 26 through 50. No reason to look through all 200 at once.

I don't see why it needs to be a silver bullet.

For example, an applicant that goes through the trouble of writing a thoughtful, reasonable README for every one of their projects signals a positive quality.

When comparing two candidates, that's certainly something to consider.

Why not? Can you elaborate? Because 10 minutes with a Github account is enough for me to decide whether I'd hire someone or not.

Why? A lot of people (if not most) treat their non-work GitHub as a place to host random experiments and projects. Unless one plans to publish projects for others to use, no one should be forced to have perfect code for a one off idea if they don't want to.

There are a lot of great engineers with empty or "bad" Github profiles.

When I look at a developer's Github profile I know that:

1. They care enough to have a Github profile

2. I can see their actual interests.

3. I know they're interested in random experiments

4. I can see if and how they treat their projects. Are they dedicated maintainers, or do they just fire off one-shot projects without ever deciding to maintain them?

And a whole lot more. A person who maintains a single shell script project with diligence is worth more to me than a thousand developers with bullshit lisp and haskell repositories.

Yes, there's a bunch of great engineers with empty or bad github profiles out there. But those engineers aren't showing me that they're great engineers. The ones with a single or few well-maintained projects do.

I highly doubt there's any reliable signal in any of these points. Someone is only going to maintain a shell script if it has users or it continues to be useful to them. But that fact is not significant regarding ability to deliver quality code. Not maintaining a project on github simply signals that the code isn't useful enough to anyone to maintain it.

Many people have their favored hiring "hacks", and they almost never actually signal for the things they purport to. It's way too easy to fall into confirmation bias or overstate one's ability to judge others.

That's fair, thanks for explaining. To me it's fine to treat it as a huge bonus if the engineer can maintain a long lasting project with diligence.

I just know a lot of people find side projects fun because they can let loose a bit and try out random stuff that may not be their "best" code.

>I just know a lot of people find side projects fun because they can let loose a bit and try out random stuff that may not be their "best" code.

This is still a huge asset. It's important for a person to want to grow and learn new stuff. When it comes to hiring a person, there's a whole lot of stuff that's important. Maintaining a long lasting project is one of them. Having an interested in fun side projects, also. It doesn't have to be good code. In fact, I have my fair share of "this project is not maintained and should not be used in production code" repositories. That's just shorthand for "this code sucks".

It's funny to see all these comments about people's side projects being "loose" and "experimental". I have a few side projects, and they tend to be very clean, well-organized, polished, because code that you write professionally for a company tends to end up being the opposite: rushed, hacked together because deadlines, full of compromises, and littered with legacy crap that the business can't get rid of.

Just out of curiosity, how many hours a week all the people you hired worked for their previous employers and for how long?

I mean, what am I supposed to look at? Unless it's obviously poorly written (e.g. no consistent style) how do you judge it?

Perhaps more tellingly I've interviewed people with seemingly impressive GitHub accounts who were just terrible. That experience leads me to believe it's easy to set up a GitHub that "looks good" without having the chops to back it up.

> I can take a 10 minute look at a designer's portfolio and get reasonably accurate insight into their style and abilities.

You can't because design isn't about style, it's not art. It's about solving problems and increasing your employers profit.

Normally people's portfolio would include their actual commercial work; it's just the way copyright has turned out in software means we can't do that.

Perhaps the best solution would be a unionized right to proper credit to developers like in the film industry.

But creatives are allowed to put work done for others on their show real this is not true for professional jobs

I agree, but the problem with this is that visual design is publicly available, and most professional code is locked behind IP and copyright laws. I can't bring my employer's code base to an interview with another company without risking a heap of legal trouble.

What if we consider software candidates like tradespeople? How do electricians and welders demonstrate their ability in an interview?

For electricians and welders, there are licenses / certification. Obtaining the license / certification demonstrates the baseline competency

Isn't this like a college degree?

Work history, references, general questions and maybe some whiteboarding. Honestly just a regular interview.

Software development is certainly more creative than electrical or welding work, and thus certifications are more meaningful. In software engineering, what certifications you have or what languages you're familiar with is a comparably poor predictor of your success on a given project.

The next time I'm interviewing in front of a hiring manager who asks me about my side coding projects, I'm going to ask them about their side managing projects.

I'm a hiring manager, and I've been dying for someone to ask me a question like that. I'm proud of the mentoring I've done, and I have a whole bunch of folks I'd love to have candidates talk to to see how I am to work for. But noone ever asks (understandably), and it feels pretentious to force the issue.

Then you will not get the job, which I thought was one of the points of job interviews.

Maybe that's not a job I want, if the hiring manager is lazy enough to 'require' side projects to consider a candidate.

> Maybe that's not a job I want, if the hiring manager is lazy enough to 'require' side projects to consider a candidate.

I understand that you might not want a job there and that part of a job interview is deciding if the company is a fit for you. But I wouldn't call a hiring manager lazy because they want to see examples of your work, in fact that's their job. How is a hiring manager suppose to make a thorough evaluation of a candidate if they don't show their work?

And how an I supposed to tell whether they're worth working for without seeing their work?

> And how an I supposed to tell whether they're worth working for without seeing their work?

Ask questions, and do you really expect a hiring manager to have work relevant to your interview to show? But you can ask about policy, culture, etc.

So ask candidates questions. You really expect candidates to have work relevant to your interview to show? But you can ask how they deal with problems, cultural preferences, etc.

> So ask candidates questions. You really expect candidates to have work relevant to your interview to show? But you can ask how they deal with problems, cultural preferences, etc.

First of all thanks for being so disingenuous with your responses. But a programmer is actually creating things ie, code and programs which can be shown. A manager is talking to people, going to meetings, emailing people, ie not much work to show, because their work boils down to talking to people. So basically different jobs have different job interview processes, you wouldn't hire a designer or architect without seeing their previous work.

I realize that you're being glib, but I have had a candidate ask me about someone I've mentored (which is basically a management side project) and I thought that was a great question.

Is this actually unrealistic? About half the managers I've had have been "working on a startup idea" or some such.

Then they'll tell you about what they do for their kids, or how they organise their local squash club or charity ... It's really not asking a lot to demonstrate transferrable skills outside of a day job.

Haha. That's called my kids and if you want them at your interview, I'll be more than happy to bring them.

Which would be a very useful answer if it suggests you view your workers as children.

yeah and they are going to write in the slack channel how you were so aggressive and extrapolate how bad of an employee you would be for the company, challenging every little thing with abrasive non sequiturs

cute, but you'll adapt.


My life drastically improved when I limited who i gave a fuck about. People who don't hire me don't make the list.

I highly recommend not giving a shit to anybody else.

I'm sure they'll do that, and I'm sure they'll be happy they didn't hire me... and likewise, I'll be happy I'm not working for them.

You won't likely get hired anywhere without a work sample. Wether that be a side project, code from previous employment, or a coding test, you'll have to demonstrate you can do the work.

Many other professions also require work samples. Even salespeople are required to produce metrics from past jobs on the number of sales they made, and get put through practicals where they need to actually do a demo pitch for something.

>Or what about other engineering disciplines? Does one really ask a Mechanical Engineer or a Civil Engineer about their side projects?

Not civil engineers (because it's hard to build a 4-lane suspension bridge on a Saturday) but very often candidates in electronic and electrical engineering have side projects that can help them land a job. Are the electronics breadboard projects that you did in your garage absolutely required to get a job?!? No, but they help you stand out among the crowd.

The reason that <computer programmer> is different from your analogies of marketers, lawyers, recruiters, etc is that the activity of programming is often a source of fun and play. Some children with curiosity start programming on their own and continue it into adulthood. (And hence, they might have some personal github projects.) In contrast, people typically do not do "lawyering" for recreational fun.

Programming is enjoyable enough for some that side projects on github are mostly a voluntary and organic phenomenon. Some employers are taking advantage of that signal and prefer to hire those types of programmers.

> In contrast, people typically do not "lawyering" for recreational fun.

Indeed, it's the opposite - industry bodies have to mandate pro bono work, which wouldn't get done otherwise.

Computing is really noisy and chaotic. There are a lot of dudes out there that just can’t cut it, for whatever reason, and there’s no good way to figure out who’s who.

Whiteboard stuff selects for a particular algorithmic dexterity that I just don’t need a lot of the time. And rejects some of the grungier virtues that make a big difference in moving units in production.

Hiring folks for a couple weeks is fantastically expensive in management overhead. I gotta be highly confident you will work out to take that step.

Just hiring a bunch of dudes and firing the duds is similarly expensive and I frankly don’t want to be that guy. I hate firing people just because I was lazy interviewing. This is my job and I take it seriously.

And there are a whole lot of guys that just don’t ship. They show up. They commit code. They are prompt and attentive at meetings. But if you need to ship, they just can’t shit it out. So many organizations are rotten with these guys- you wonder how so many devs can produce so little product, this is the reason. (Also shitty management- mgmt is ten times noisier and a hundred times more expensive if you fuck up.)

> And there are a whole lot of guys that just don’t ship. They show up. They commit code. They are prompt and attentive at meetings. But if you need to ship, they just can’t shit it out.

Inability to ship tends to be a project planning and scoping problem, not an engineering performance problem. If you can't ship, it's not because you have lazy developers but because you either bit off more scope than you could chew, or you don't actually have a plan that gets you from-here-to-shipping--one that is more detailed than "work harder, guys".

I'm kind of a slacker responding to HN stuff, but here's what I'll say:

Yeah, you're right, if shit isn't shipping, it's my fault.


There are a variety of psychological profiles in computing, and they have various pluses and minuses. I need folks with kind of a... move fast and break things attitude. That's what I select for.

Furthermore, the 80/20 rule is a super real thing and the last 1% until a project ships hurts real, real, real bad. There are people who can push through that and there are people who can't. Yeah, it's my fault we're not sipping tea as we ship, but I've been doing this a long time and I'm pretty good at it, and I can't make that pain stop. I just can't. I've been on teams that shipped and suffered, and I've been on teams that didn't ship.

If somebody's cracked this code for me, man, throw me a bone. But while I'm always trying to do better, I also kind of have given up. I'll make it up to you on the other side, I promise, but it's gonna hurt a little pushing it out.

>Does one really ask a Mechanical Engineer or a Civil Engineer about their side projects?

Yes, sometimes. I'm a mechanical engineer, and frequently interviewers would ask what sorts of projects or engineering related hobbies I have. They're not necessarily looking for me to have brought a mass-produced consumer product to market in my spare time... most mechanical engineers I've met seem to be excited enough if you talk about working on cars or bikes.

I can see how software engineering may be similar. They may not be expecting you to be working on some hugely popular app, they may just be trying to get a gauge on what you're really passionate about. I don't like to wrench on cars in my spare time, but I've still gotten job offers after it comes up. In the end it's just a matter of finding the team that has similar values to you.

Asking for side projects is essentially saying that you think the resume is a pack of lies from back to front.

My work is my work, and my play is my play, and I question the wisdom of judging the play in the decision to hire someone to work for you. Even when I do code at home, away from the office, I do it differently. It might be in a different language, for a different target platform, for instance. I might write more automated tests, because I can't hand it off to a tester that isn't also me. I might write something with a command-line console interface instead of a GUI, because my user base is probably just me, and I know that person is not a helpless idiot who can't read a man page--a man page that I probably will never write in the first place.

Real work is reading the crap code written by your predecessor, understanding it just enough to change something without breaking everything else, and moving on to the next thing as quickly as possible. No interview has ever tested my ability to read existing code or to alter existing code for a specific goal.

Whiteboard this: "This third-party data grid inexplicably causes all text typed into this column to come out backwards. You have 60 minutes to fix the problem, because code freeze is in four days, test needs at least a week to bang on a release candidate build, we ship at the end of the month, and you have 15 more equally-stupid issues waiting for you." And then you are judged on the quality of your horrendous kludge. Those companies that look for perfect people that write perfect code at work, then go home and keep doing it, are living in Cloudcuckooland.

Right now, my side project is getting the last two achievements in Fallout 4. After that, it will be banging out an entire novel in November. Or maybe it will be making furniture to hold DVDs, with a built-in system to deliver a painful shock whenever someone puts them back out of the correct order (which is sorted first by genre, then alphabetical by series name or title). Or maybe it will be working out a plan to terraform Venus using wormholes. When I leave work, I do whatever I want. I don't want to endlessly prep for an unending parade of stupid tech interviews.

This is a great idea! I would love to have an interview where I was presented with some terrible code and asked how I would refactor. That would be so much more relevant to actual work experience than solving technical puzzles.

> It's funny that the tech industry is so insistent on side projects. I mean - would you hire a marketing person based on their 'side projects' or a corporate lawyer purely based on their pro-bono work?

I guess our field is so enjoyable that many people have side projects. Apparently, some recruiters think it's a good indicator of future success. I wonder if it's the case though. I teach CS and usually, passionate students (those who work on their own projects) are indeed better and spend much more time working on their assignments. But it's not the end of the story. For example, some are only motivated in things that involve coding, are stubborn, or are bad team players! Conversely, there are students who aren't hobbyist programmers but are smart, learn fast and are a joy to work with.

You know what's universally hard? Finding people who are good at their jobs. References don't matter, experience doesn't matter, education doesn't matter. Why not? Because it doesn't say anything about actual skill. What does? Side projects, where you can actually see if a person has skills.

> a corporate lawyer purely based on their pro-bono work?

You hire a corporate lawyer based on their public reputation. For a programmer, that's their github portfolio.

Business are in a unique position to actually judge programmers by their merit, not just hearsay or bullshit references or lies on a resume. Of course companies are going to factor that into their decision!

Some of us have NDAs as part of our job.

Not everyone is happily doing FOSS projects.

I think the big misconception is that a side project has to be a prolific open source project. As someone who has interviewed developers, someone who has _anything_ on a GitHub page stands out in the crowd. Even just a folder with a few batch scripts can go a long way.

Or even just a webapp where you can say "I can't show you the code, but I made it [look, my name's at the bottom!]".

> You hire a corporate lawyer based on their public reputation.

But that reputation comes from his actual daily work. I as a developper don't get paid to improve my github portfolio all day long...

a lawyers public reputation is built through the day-to-day work. Not a programmers can put their body of work on public display

Except that, side projects are nothing like day to day coding in a company. Beyond basic coding skills, they are very different animals.

I use side projects to figure out what aspects of software engineering the candidate seems to enjoy. Do they fiddle with the CSS nonstop to get the margins just right? Are they constantly trying to decrease the memory impact of a function? Does the README read like an appliance manual or a haiku of puns?

Side projects are a strong signal the candidate enjoys this line of work, and ideally I'd want them working on things they would enjoy.

I would prefer if you would ask me about my preferences then guessing them like that. You would got them wrong.

I would also prefer if I did not had to worry about readme or css projecting this or that image.

Actually, the greatest predictor of employee success is their IQ, but that's illegal to test. That's why some employers substitute SAT, ACT, ASVAB, etc. scores, since those are abstractions to IQ tests.



>A growing body of research suggests general cognitive ability may be the best predictor of job performance.

In my experience the industry is not at all insistent on side projects; lots of people talk about having them, and the way people talk about them has become weirdly culty in the last five-ish years, but it is entirely one-sided; nobody who is doing the hiring cares at all.

Developers are more akin to musicians or artists than a marketing person. Yes, I do expect musicians to practice their instrument and learn new material in their free time, at least a little bit. A musician who _only_ played during paid gigs and never practiced 1 hour outside of the paid jobs, would never learn anything new. They would forever be stuck with their existing repertoire.

Would you hire a musicians who didn't have recordings of their work and wasn't willing to perform an audition? Would you want to hire a musician who didn't practice new things?

The development landscape changes rapidly. Frameworks come and go. If you want to keep up in the industry, you will most likely have to learn a new language or learn a new framework at some point along the way. If you only code during your 9-5 you are forever stuck with your current skill set and will end up like the people in 2017 who only know COBOL and can't do anything else. Your job options will shrivel and you will not be very marketable.

I think it's fair to consider how passionate a person is about improving their career skills. For programmers, that often consists of side projects. For other professions, it could include going to conferences, taking classes, reading trade publications, etc., all of which would be worth mentioning in an interview.

I don't think it is fair to imply that people who don't have side projects aren't passionate about improving their career skills. I don't have a lot of side projects, and the ones I do have were hastily hacked together out of boredom or convenience and aren't anywhere near the quality I'd want to show to a potential employer, but I do lots of other stuff to improve my skills, like reading books, listening to podcasts, attending conferences, etc. All of these things improve my career skills, but the improvement is primarily demonstrated at work not through a bunch of side projects.

> It's funny that the tech industry is so insistent on side projects.

Are they? I've yet to be asked about it. It might be different for new grads, but what people seem to care about most is if you've worked for a major firm, or in the case of the video game industry, a "AAA title".

This is like a doctor saying "It's funny the medical industry is so insistent on medical licenses. I mean - would you hire a software engineer based on their 'license'...".

Different things are different.

Is software development a science or an art? Most artists have portfolios.

Most artists own their own work and are not under NDAs.

Graphic Designers, Copywriters and Art Directors ARE all supposed to have side projects AKA a portfolio.


Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact