IRC as a Hiring Filter
jasonwhaley.com
jasonwhaley.com
Recruiting people you find on irc in coding related channels I'm all for that. But it doesn't scale obviously.
And yes, make senior people write FizzBuzz on a whiteboard, they find it thrilling.
I've told potential employees "no" more than once due to sequences of hurdles during the hiring process that were easy to pass, but told me very clearly I didn't want to work there.
The more bullshit, the more likely it also is I'll receive and say yes to another offer while they're busy playing.
Yet I'm not on irc, I'd be a hire you'd miss if you used irc as a screening filter.
> "and I have actually decided not to work with specific developers before in a professional setting because they literally could not figure out how connect to freenode and join a keyed channel"
How about you just talk to people in the way they're used too? Instead of deriding them for not using <insert favorite means of communication here>. I know how to use irc, so this scenario does not completely apply to me, but it's very possible that a very technical person has never seen it before and has trouble with the irc commands or w/e. Especially since irc is really becoming a relic of the past, and newer hackers/programmers might never have used or seen it.
technical person having trouble with the irc commands - isn't that oxymoron?
Irssi will lift your nickname from the username. It may even come with freenode pre-configured.
/ignore #room-name JOINS PARTS QUITS
Can you imagine that it'd be kind of hard to join to join a channel for your interview if you've never been told how or were granted the time to find out?
Sure it's popular in opensource/hackers/etc, but I'm sure that 'people who use irc' is a true subset of 'people who are technical' and not the other way around.
But what if you're, as the article says, used to skype?
would it occur to you that there might be web based irc? Skype doesn't support that, and certainly the only things you see when you google 'irc' are clients you'd have to install, or worse: compile. Tough freenode does seem to have the webclient high in the list of results...
But then looking at their client is another example of why irc might not be straight-forward enough to use. This is what happened when I tried to use it with a random, but obviously too short, user name:
--- [19:43] Licensed under the GNU General Public License, Version 2.
[19:43] == * (qwebirc) Looking up your hostname...
[19:43] == * (qwebirc) Found your hostname.
[19:43] == Connected to server.
[19:43] -herbert.freenode.net- * Looking up your hostname...
[19:43] -herbert.freenode.net- * Checking Ident
[19:43] -herbert.freenode.net- * No Ident response
[19:43] -herbert.freenode.net- * Couldn't look up your hostname
[19:43] == CGI:IRC host/IP set to gateway/web/freenode/session ...
[19:43] == bob Nick/channel is temporarily unavailable
[19:43] == bob's Erroneous Nickname
[19:44] == ERROR: Closing Link: gateway/web/freenode/session (Connection timed out)
[19:44] == Disconnected from server: Connection to IRC server lost. ---
Riiigiht, very usefull, I can't use the nick bob. What now? How do I change it? where's my change button? etc. etc. I think my point stands, irc might not be 'hard' but for a first time user it's not easy either.
That makes little sense to me, if the point was to use the applicants ability lean how to use a new technology as a screening filter then I don't see how using irc fits that requirement.
Although I'm opposed to the idea of using irc as a screen, since I know that there are people who don't use it but are technical, I'm not suggesting that most technical people don't know what irc is. Just that it's kind of useless to use a known technology to test the ability to learn new technology, especially amongst developers.
Further, he never mentions exactly how long they gave those developers to figure it out.
The post is a bit unclear about this, but if you give them the benefit of the doubt, then they're testing exactly for the key ability any IT worker needs: The ability to research a new technology. You make the assumption that they did not give them much time, but consider for yourself the following please.
You send a senior developer an email that says the following:
use IRC to connect to freenode,
join the channel #blah
using the password sekrit
Assuming their network does not actively hinder IRC use, what does that tell you about their ability to do basic research?
Importantly, an old technology that's new to them. This isn't researching node.js in 2012 -- there are reams upon reams of how-tos and so forth about IRC.
(in case it wasn't clear, I'm agreeing with you)
Which is not what the article says you should use this screening tactic for, since there is a large group of technical people out there that is irc-literate. Which implies that irc was not chosen, as you suggest to figure-out a persons autodidactic abilities, but rather because the author (kind-of-correctly) assumes that most technical people have at least a passing-relation with irc.
I spend a lot of time communicating with people online, but IRC isn't a channel I use. The problem is that I have a family that naturally consumes a lot of my time. That means I don't control when I'm available to communicate. I get a decent amount of total time, but it's in bits and pieces spread throughout the day (mostly early morning or at night).
That's perfectly fine for asynchronous communication like email, SO, here, etc. But it doesn't work for synchronous things like IRC.
So, using "are they on IRC" as a hiring signal, I believe, effectively punishes people who have families, busy work schedules, or otherwise complex lives. In other words, I think it ends up favoring affluent young people. That's not something I'd personally be keen on biasing for in my hiring process given how much of an advantage that group already has in technical jobs.
From a hiring manager's standpoint, isn't more free time a benefit?
The same thing could go for using GitHub code as a signal. People with no free time are less likely to contribute to the community there too.
Lack of Github means I'll have to find another way of seeing example of the persons code, but that's ok. I wouldn't dismiss anyone as a candidate for that.
But someone with a Github repository or ten might be at an advantage because it serves effectively as lowering the threshold: I don't need to find their CV interesting enough to decide to ask them for an example of their code, just interesting enough to look at what's already there.
IRC is highly asynchronous. It makes it possible to communicate at near-synchronous speeds, however at the base of the technology, any semi-competent developer i work with, even those with 3 kids, are able use IRC consistently, with the only impact being that answer latencies range between seconds and days. Then again, i have the same freedom, so it benefits both sides.
For a while, it was my main source of social interaction overall. 20 years ago I used to pick up women on IRC - most of my dates for years came that way.
At the same time, while certainly there are women and minorities on IRC, I've seen few environments that could turn incredibly ugly and sexist and bigoted so quickly, or that could wear women in particular down so quickly and drive them away. It at least used to take an incredibly thick skin to be a woman on IRC.
In fact, back in the day, one of the most effective way of picking up women on IRC, was to simply set yourself apart from what they quickly became used to, by making fun of the endless stream of horny teenage boys anyone with a somewhat female-sounding nick used to have to deal with, no matter what channels they were on.
The extent certainly does vary by network, and channel, but watching over the shoulder of a woman using IRC anywhere without very strict policies and very specific audiences can be very enlightening unless they hide their gender. And many do.
No hard feelings. I lived through a fair share of flamewars, and stuff like karma holds little meaning to my overall well being.
"while certainly there are women and minorities on IRC, I've seen few environments that could turn incredibly ugly and sexist and bigoted so quickly, or that could wear women in particular down so quickly and drive them away."
Here is the whole problem - claiming that IRC is bad for minorities is like claiming that all establishments serving food and drinks are.. and while this may be true for pubs, it certainly is not for restaurants.
It all depends where you hang out, and on technical channels, I have never seen anything that would suggest that minorities are not welcomed.. in fact nobody cares who or what you are, and you have no reason to share that information if you believe that this somehow may affect how people treat you.
This article is a pretty solid point of evidence that IRC communities are insular against every non-member, regardless of who the non-member is.
shib·bo·leth
SHibəliTH,-ˌleTH
noun
a custom, principle, or belief distinguishing a particular class or group of people, especially a long-standing one regarded as outmoded or no longer important.
"the party began to break with the shibboleths of the left"
But as a general practice, I don't see why this would be more beneficial than presenting a candidate with a rotary telephone and telling them to make a phone call with it.
I felt terribly old the other day, as one of my younger co-workers overhead a conversation where rotary telephones were mentioned and turned around to ask us "what is a rotary telephone?"
Then again, I don't think my 6 year old son has actually seen a wired telephone yet (we do have a land line, but it's wireless too) before, so I'm sure said co-worker will get the pleasure of feeling old soon enough.
IRC is still useful and relevant for a lot of OSS, but it's definitely not something relevant for screening people.
However, if you wanted to create an easier to use version of IRC/git/whatever, would that be easier if you didn't understand the original or if you did? I'm on the fence about this. I'm glad I remember one thing that confused me about git, although I probably forgot others.
* No one in your notify list is on IRC
Something I find really useful is Jira integration with IRC. When something goes wrong, an IRC channel will open up with the name of the Jira ticket -- then status bots and logging bots will join automatically. It really helps to organize everything, especially when multiple problems are happening concurrently. It's also great for building post-mortem timelines after things are resolved.
In addition, most teams also have their own channel, so you can generally join it and ask questions for a quick response.
I built Slack integrations like this for ongoing development, but your use case for troubleshooting sounds amazing. I would love it if you wrote a blog post about it.
I've heard someone comment to the effect of "if you see a netsplit when you're in a war zone, prepare for the worst"
Over 100 or so already idling in the channel.
I'll see you all in there!
It's a pretty easily pattern to fall into - people who work like I do must be pretty smart because I try to do what works best. It's hard to evaluate someone whose mindset and approaches are different than your own. But when you've been around for a while, you realize there are so many different ways of solving problems and areas of expertise that there's no one right answer.
So ultimately, I don't think these kinds of filters are that useful. It's basically a yes or no question, where a "yes" reinforces your existing biases and a "no" doesn't really gain you any information.
If you gave someone time to figure it out, and they seemed unwilling or unable to figure out how to get on IRC, then yes, that'd indicate someone unwilling/unable to figure out a relatively simple tool.
But expecting them to already (or still) be IRC users today is a bit like expecting them to still spend lots of time on USENET.
Looking at the drastic decline in user numbers on the major IRC networks as numbers of internet users have skyrocketed should be enough to make it clear going by irc usage is likely to be a stupid idea.
Then again, I'm biased - I've not used IRC seriously for a decade at least, and did most of my IRC usage (including customizing irc-ii) about 20 years ago.
---
A custom, principle, or belief distinguishing a particular class or group of people, especially a long-standing one regarded as outmoded or no longer important.
---
Besides the mentioned Github and open source project participation other shibboleth I use is some experience or knowledge of functional languages -- Scala, Haskell, Lisp, Erlang, F# and such.
I don't use them as hard guidelines but these just proved to be good heuristics in the past. Chances are high if someone is good at Haskell they will be good at many hard problems we throw at them.
But, I think what you're trying to hire are skilled software developers. It's apparently been left as an exercise for the reader to demonstrate any correlation between these two skills.
Filter does not mean barrier.
Kids these days can't use IRC, don't know that gopher still exists, never used usenet...
Being a software developer, or admin, or even technical writter, and never even trying to learn how the Internet looked like before everything was turned to Javascript...How sad..
It's like never riding the bike...
insightful? it tells you if the person uses irc on a regular basis already (in which case they might have a native client) or whether they use something else usually (in which case they might use the simplest web-based client they can find on google).
I've personally struggled with this; as a young teenager, I tried to learn C++ without any access to the internet, any mentor; just a few old books. I struggled with it for years and ultimately failed because certain technical obstacles couldn't be overcome without "in-group" knowledge that didn't exist in the meager documentation I could find. Either you were part of the social group that could help you out of your problem, or you effectively were not privileged to use the technology.
IRC is symbolic to me of this sort of old-school "it works fine for us" mentality that poisoned the early tech enclaves - their myopic failure to include anyone who didn't win the social lottery of knowing the right people, or growing up in the right place. It's not that it worked for them that I begrudge - it's their assumption that things must also be working for the rest of humanity, since they themselves were doing just fine. It's the same sentiment that made people hate the classic "let them eat cake" line.
(As a corollary, this is why Stack Exchange is maybe the best thing to happen to tech since the internet - a system specifically designed to incentivize people to document all of those "undocumented secrets" that are necessary to actually get any work done in a given field.)
Finding sensitivities to diversities of technical literacies is troublesome but teachingsome.
That means identity issues are social issues, and language choices, changes, modus, and tools are important to address if we are serious about giving all forms of identity a chance.
For example, I usually only register my nickname with nickserv when I know I'm going to be a (semi-)regular user of network X. Same with setting up my BNC to connect to that network.
It's true that we sadly don't have a significantly better alternative, but that doesn't mean one can't have good reasons to avoid it.
Let me know if you have a position open for a Gopher developer.
It's great you mention 'being a prick,' because that's what I said out loud as I read your article. I was tickled by the fact that you mentioned it, so I do respect that you know your strengths.
This might work for you, and for the projects you work on, and I think it's not a bad recommendation for positions that need... that kind of stuff.
That said, experience causing net splits by CTCP flooding #quebec on undernet was helpful for understanding a situation where Erlang/OTP pg2 was flooding itself into oblivion. :)