Fog Creek Software
Discussion Board




The Guerrilla Guide to Interviewing

http://www.joelonsoftware.com/articles/fog0000000073.html

Don't you think Joel's "The Challenge" technique is forever spoiled for candidates who happen to be regular readers of joelonsoftware.com? Me, I'd quietly lay an extra layer of stubborness upon myself if he happened to start passionately defending weird opinions during the interview.

Ah, and by the way I am one of the Chosen Few who GETS POINTERS!!! Bow before me.

Klodd the Insensitive
Thursday, November 06, 2003

Who cares if they've read JOS?

I've interviewed people who wouldn't disagree with me, even if I stuck a gun to their head and told them the sky was pink. If a candidate has read the guide and has learned something, then mission accomplished.

anon
Thursday, November 06, 2003

I would not disagree with you either, if you held a gun to my head...

Practical Geezer
Thursday, November 06, 2003

#6 is a trick question.

I have put a lot of effort & thinking into designing houses that appeal to the mass market in the midwest.  I've thought and discussed with colleagues things like recommended amount of lighting per square foot, triangle design, color schemes, cubic feet of storage, ...

Let me tell you son, the business of business is to make money, and I know how to design a house that will do that.  If you wouldn't hire me because I started out with a square on the whiteboard, then you're just playing semantic games and I've got no use for you.

Mr. Russel, contractor turned software inventor
Thursday, November 06, 2003

"I would not disagree with you either, if you held a gun to my head... "

Heh, heh -- funniest comment I've read here all week.

J. D. Trollinger
Thursday, November 06, 2003

I've designed a few houses myself and if you're the same type of developer as you were a wood butcher... good riddance.

Air Marshal Carlin
Thursday, November 06, 2003

[I would not disagree with you either, if you held a gun to my head... ]

Doh! I knew that didn't come out right...

But I might have hit upon a way to reach a quick consensus in our meetings.

anon
Thursday, November 06, 2003

I'd hold some personal reservations about the actual gun part, tho - does that count?

Philo

Philo
Thursday, November 06, 2003

Like most simplistic rules, I take issue with this:

"Which brings us to #7, The Challenge. This is fun. Throughout the interview, you look for the candidate to say something that is absolutely, positively, unarguably correct. Then you say, "wait a minute, wait a minute," and spend about 2 minutes playing devil's advocate. Argue with them when you are sure they are right.

Weak candidates will give in. No Hire. "

Really? What if you accidentally, in playing devil's advocate, give them an aspect they hadn't thought about, invalidating their position?

For example, it's 1994 and they mention how all libraries should be COM .dll's, since it eliminates versioning issues. You take the contrary position, and something you say makes them realize that if someone else accidentally releases an earlier core dll with a later version number, you could wreak havoc on the IT world. So the candidate realizes that COM is a house of cards and maybe we should find another solution, like digital signing of code libraries...

You're going to "No Hire" the guy who discovered dll hell before MS and accidentally invented the .Net Strong Naming convention?

[I'm fairly certain that Joel puts more thought into it than a simple "oops, you changed your mind, you're out of here", but I'm taking issue with the absoluteness with which he states the rule]

Philo

Philo
Thursday, November 06, 2003

Dear Contractor,
                          I wouldn't hire you if you started with a square because the plot of land I've got needs a rectangle. I've just spent countless hours designing my own house and sending the building plans off.

                          What makes you think the house is in the midwest? Or that you are going to make any money trying to sell your mass market design to Joel's mate, the investment banker.

                        The question is designed to weed out people who jump to conclusions, which is what you just did.

Stephen Jones
Thursday, November 06, 2003

Don't be a dumbass Stephen, you're talking about a custom house on an odd lot and I'm talking about developing the mainstay of a product line.

Mr. Russel, contractor turned software inventor
Thursday, November 06, 2003

If you asked me to design a house, I would automatically assume you were talking about designing my own house.

Why?  Because that's what I'm doing NOW!!!

I guess I'd lose, but this is all I've been able to think about for the last several months. :)


Thursday, November 06, 2003

You wouldn't have been able to hire me anyway; I only worked with land developers in the mid market range.

See, you're doing just what I did:  Jumping to conclusions. 
The question was 'design a house', not 'design a house for Stephen'.

But, you are right, I wouldn't build without a spec, and I am just being a little onery today.

Mr. Russel, contractor turned software inventor
Thursday, November 06, 2003

Well as far as Mr. Russell is concerned you'd lose because he would presume that you must design a house for an undefined product line in an area he hasn't told you about for a client you don't know; and if you can't read his mind you're a dumbass.

The point of course Mr. Russell is that you don't know what kind of house the person who asked you to design one wants, so you are wasting time if you go ahead without asking for more details.

And where I'm building you wouldn't make much money with square houses as a product line because most houses are rectangular.

Blank, I sympathize with you :) I know the feeling. But if you are so pre-occupied with your house do you think you would be a good hire :)

Stephen Jones
Thursday, November 06, 2003

One shouldn't design a house by initially drawing a square because the customer always knows exactly what he or she wants, and the use of visual aids to elucidate feedback is just plain silly.

Devil's Advocate
Thursday, November 06, 2003

I live in a square house.

Never make sweeping generalizations about anything.

Rob Van Hoose
Thursday, November 06, 2003

I prefer those extra dimensions...  easier to put away the groceries.

Air Marshal Carlin
Thursday, November 06, 2003


The point is, we all operate under certain assumptions.

Joel is operating under the assumption that the programmer wasn't involved in the house construction industry.

Open your minds, and flip it around a little bit.

Imagine if you asked a Doctor how he would design a clinic.  He probably has done this in his mind a million times, and has a pretty good idea what it should look like.  He's lived & breathed this world for years.  He knows what he wants.

Trick.  I meant a clinic to treat sick reptiles located in Guatemala.

Janome
Thursday, November 06, 2003

I hate playing games with recruits.

My general preference for checking out personality and assertiveness is to either ask them a really really hard problem that they "should" know but generally don't or see how they react to being told that they are wrong.

The first is great.  My best ones have been the really hard questions that can be scaled back to something that somebody will know.  Watching to see if they give up, grab for a lifeline, get an attitude problem and start talking back to you (that one's a definate classic horror story for me), or what is quite instructive about how they handle the hard stuff.

The second one is great, too.  Especially if somebody seems a little egocentric, if they are egomaniacal to the point where they can't handle being wrong, I don't want to work with them.

None of these require outright lying to the recruit, which in my mind is a little unfair.

I think the best story was that somebody pointed out that he was in an interview where he had to argue against a completely absurd argument.  Except they weren't pulling a Joel, they were off their rocker and a horrible person to work under.  So, in a tight market, you may lose an otherwise good candidate because they are sure that you are off your rocker.

I think it's also the case that you will generally be able to tell if somebody has the "key" to your interview, no matter what. 

Flamebait Sr.
Thursday, November 06, 2003

Flamebait Sr, could you tell the horror story?  Those are always interesting...j

tester
Thursday, November 06, 2003

There's not much to tell without getting into details I can't really discuss.  If it's a tad verbosely vauge, sorry, that's all you get.

Person was in for interview.  Flamebait Sr and another engineer one year his junior do the honors for one section.

We began the interview, they were doing OK by our standards although this section of the interview tends to cause about 50% of our successful candidates to think there's no way they got the job because it's hard.  It's a long drawn-out problem where we give them some time alone with it and then go through the whole thing together.  We often times need to point out specific parts of said problem because next to nobody gets every little detail.  As in, it's not expected that they ace it. 

As time goes on, they start to get more and more visibly frustrated.  It's not that they are not *getting* the details, the lights will come on, but they are getting frustrated over not noticing them.  With other vict.. er... candidates, it's often how fast the light comes on that makes the difference.

They start talking back, asking us how we could expect them to have noticed said detail.

It's frustrating for us because if they were obviously a wrong fit, we'd just be easy on them, get it over with, and move on without crushing their spirit.  We don't do that unless we've made a no-hire decision.  But they are talking back and it's getting to be so tortured, we try to get through the whole thing without being too hard.

They also complains that they have a learning dissability, they ask how the interview went, and things like that.  They complain that it's not a fair interview *and* that the helpful-but-vague answers I gave them in response to their questions about said problem steered them wrong.

For comparison, it should be noted that Sir Senior Flamebait also has a learning disability.  I do not expect, therefore I do not extend, any special treatment, positive or negative because of that.  I was overall annoyed that they brought it up because they *know* that such things are sticky legal issues and not allowed to play in to our decisions.  If facial, vocal, and body-language clues weren't important to communication, I'd half want to interview people behind a screen and with a voice scrambler so that I could assure myself that I was not showing bias.  If there's something we should know, let us know after you've started.

Towards the end, we've totally slacked off.  It's considered rude to just stop the interview and kick the person out then and there, so we're just trying to bring things to the point where they get the feeling they got a fair shake.  And the candidate is so frustrated that they can't answer even simple stuff at this point. 

The wrap-up is a very insightful end to this story.  We declared it the worst interview ever (and it has yet to be topped).  Person came recomended by existing employee.  Other interviews were either OK or good.  Because of our interview session, it reflected badly on their personality (When stuff is crashing and nobody knows why, would *you* want somebody who got that frustrated in a stress situation? ) and they were a "no-hire"  Were they to have been able to keep their cool and not talk back, I believe that they would have ended up high enough in the ranking to have been extended an offer, albeit not as one of the top candidates.

It was very depressing and draining.  Normally after an interview it's either neutral, a feeling of "Cool, I hope we hire this person, I like them", or some snickering while the candidate is out of earshot.  This was just depressing because I could sense that the candidate had at least adequate raw intelligence but they just needed to apply it and quit it with the attitude.

Flamebait Sr.
Thursday, November 06, 2003

No, I think I can see where Mr Russell is coming from.

What he is saying is that if the house is a euphemism for a software package, you know what the function is (it is for someone/thing to live in, that is what houses are for). You don't know the target "audience", but it /should/ be "the audience that makes me/us the most money" (taking into acount size of possible market, competition, company capacity, etc.). If your business plan is to sell houses to giraffes then good luck to you, but I don't actually think I want to work for that company because it will not be profitable. The same applies if you want me to create a new OS - the probability of scoring big when set against the existing competition is so small that you are better putting your efforts into something else, otherwise you end up like Apple or BeOS.

That's not to say that competition against the big boys isn't possible, but you have to pick your battlefields and you'd better have some compelling reason why customers should choose your houses over, say, Microsoft's. Things change a lot, of course, if you are building houses for free.


Friday, November 07, 2003

So, if someone said "design me a computer program" you'd rush straight off and start coding?

Stephen Jones
Friday, November 07, 2003

But who would say "Design me a computer program!" in the real world.

They'd talk about something they were already doing manually (say) and ask you to "automate it".

Or they'd tell you that Internet Explorer Sux and they want you to write "a better browser".

Now in either case thats not enough to roll up your sleeves and code but it does give you a starting point.

Robert Moir
Friday, November 07, 2003

No, because you don't know the function of it. I pointed out in almost the first sentence of the other post that the function of a house is known. If the question was about designing a building it would be different.


Friday, November 07, 2003

The prime function of a house (to provide shelter) is known but that is not sufficient to start the design.

Shelter from what, where, how big, how expensive, what terrrain, what materials, for how many people, what secondary functions, etc, etc, etc.

Stephen Jones
Friday, November 07, 2003

"ask candidates to design a house"

is a lot different from

"ask candidates to design a house for a client"

or

"ask candidates to design a house for 40 foot giraffes"

Pretty stupid question.  Much better would be "how would you go about designing a program that did X"

If I had a Mr. Russell in my office, and asked him about the perfect house, I would expect him to draw on his experiences and tell me what he considered a perfect house.  And if he mentioned something about a house that makes the most money, he'd be hired on the spot.

Johnny Simmson
Friday, November 07, 2003

---"Pretty stupid question."----

That's the whole point Joel is making. It's a stupid question and if you go about trying to answer it without clarification you will be a very expensive liability to any company that hires you.

Stephen Jones
Friday, November 07, 2003

Stephen, do you really think that?  Do you really think that if the guy above that is building his own house went into detail about designing a house that he is somehow a lousy programmer that will cost the company money?

I think that's an incredibly naive view. 

Johnny Simmson
Friday, November 07, 2003

What costs the money is making the assumptions.

Now if he'd said, "well, actually I'm in the process of designing a house at the moment, so I'll tell you about it" he's not making assumptions.

A simple school attendance database has a field for absences. They didn't ask any questions and made the field Boolean. "He's either there or not. right?" Wrong - he could be late, and that counts as one third absent. Nasty rewrite necessary.

There's nothing necessarily wrong with making assumptions - as long as you know you're making them.

Now, this question is not aimed at people who have designed houses. If you knew they had, you would ask them to design a car. However it does seem a good way to tell if a candidate views design as something absolute, or the result of a series of trade-offs made on the basis of detailed solicited input.

Stephen Jones
Friday, November 07, 2003

Scene: Job Interview with Joel.

Joel: Can you design a house?

Me: Sure. (goes over to whiteboard, draws box with trapezoidal roof, little door, two windows, and a chimney with smoke coming out).

Joel: (slightly annoyed). Actually, I meant a house for a family of 40' high blind giraffes.

Me. OK. (draws a scale next to the house showing it's about 60' high, draws foam padding round the front door, and surrounds the flower with a protective guard).

Hire or no hire?

Mr Avaya
Friday, November 07, 2003

Joel: I meant a house with Giraffes who have cubilophobia (fear of squares). 

bob
Friday, November 07, 2003

You would be a definite Hire.  But only if you lit up a cigarette after drawing the foam padding.


Friday, November 07, 2003

"BUT, good candidates will tend to get fairly passionate about the argument, and they may momentarily forget that they are in an interview, and they will get very involved in trying to convince you. These are the people we want to hire."

Or the candidate will think you're an idiot and decide you're not the boss they want to work for.

"Ask the candidate to design something. Jabe Blumenthal, the original designer of Excel, liked to ask candidates to design a house. According to Jabe, he's had candidates who would go up to the whiteboard and immediately draw a square. A square! These were immediate No Hires. In design questions, what are you looking for?"

Or these people might be more visually oriented, and even if you end up with a round hobbit hole, they like to have the square as a starting point, a common ground to work from.

"Look for closure. This is part of Get Things Done. Sometimes candidates will drift back and forth, unable to make a decision, or they will try to avoid hard questions. Sometimes they will leave difficult decisions unanswered and try to move on. Not good."

I strongly agree with this one.

----

While you can argue the semantics of the 8 step interview plan, like I just did with the square house thing, these aren't rules, it's a plan. Variation from the plan is acceptable. If he goes up to the white board and draws a square and THEN launches into some very pertinent questions, that doesn't mean you should rule that candidate out just because they drew a square.

These are heuristics, simple decision making tools, and even if you lose one or two good candidates because you crossed off the guy who drew a square, that doesn't invalidate the plan. (Note the 'typical plan' - tailor as needed.)

Joel wrote these for everyone because he interviewed a number of people, read up on interviewing, and found that this method works.

www.MarkTAW.com
Friday, November 07, 2003

*  Recent Topics

*  Fog Creek Home