Fog Creek Software
Discussion Board




Stupid specific interview questions

The candidate says they've lived their lives in New York City. The interviewer asks:
"What street is the Empire State Building on?"

The candidate says they don't know, so the interviewer ends the interview - no hire.

So the candidate has never had cause to look up the street address for the Empire State Building. They know how to get there, they know how to get to the observation deck, they know the quickest way to *find out* what street it's on. They just never happened to notice it's on Fifth Avenue.

The next candidate has been to NYC once or twice as a tourist, so of course they looked up the address and can pass your little quiz.

My point is - do you really want to hire someone based on the little specific pinpoints of knowledge you can ask in ten minutes? Wouldn't it be better to ask questions to determine how they think? Ask them to solve problems, or how they would go about solving them? Ask them what problems they faced in previous assignments and how they fixed them?

In the printf example, if you have a burning need to ask it, if the candidate says "no, it won't compile," why not talk them through it? Ask them why they think it wouldn't compile. Ask if they can think of any way it might compile.

Or maybe ask yourself if the five minutes it would take to teach the esoterics of the printf statement and string token replacement is really a reason to can an applicant.

Philo

Wisdom trumps knowledge
Sunday, November 30, 2003

Besides, how much would it cost to set up a computer, and ask people to debug or write a little proggie within an hour in a language the company asked for?

One of the companies I worked for would ask each new hiree to write a NotePad-like aplication using the company's proprietary toolkit within their trial period (3 months), to check their knowledge of C++ and their capacity to learn a new toolkit.

Combined with tech interviews with several senior developers, it was obviously a very good way to do things, since only one developer was let go during the two years I was there (and they hired quite a bunch of bodies.)

Frederic Faure
Sunday, November 30, 2003

The candidate says they have a PhD in mathematics. The interviewer asks:
"What is two plus two?"

The candidate says they don't know, so the interviewer ends the interview - no hire.

So the candidate has never had cause to memorize the answer to two plus plus.  They know how to count on their fingers, they know how to use a calculator. They just never happened to memorize two plus two.

SomeBody
Sunday, November 30, 2003

I'm going out on a limb here but I wouldn't hire anyone who didn't know what 2+2 was. And I wouldn't work at a company that asked me what 2+2 was in the interview.

;)

Matthew Lock
Monday, December 01, 2003

Somebody, the difference between Philo's example and yours is fairly simple: Although it would be easy for a New Yorker to make a, probably very close, educated guess about which Ave the Empire State building is on, the New Yorker might not 'know'. And from within the confines of the interview, there is no way to figure out for sure. However, even a PhD in math can calcuate the answer of 2 + 2 for sure.

Byron
Monday, December 01, 2003

That is a great analogy for interview criteria, Philo. The New Yorker knows how to get there, how to direct taxi drivers, everything about it. What alternative routes work if they're needed. What public transport systems to take.

Yet the New Yorker might indeed not be able to answer what others might consider a simple question - what's the address. A foolish interviewer would indeed conclude the candidate doesn't know the subject. Foolish companies would end up hiring taxi-drivers. And so many of them do.

It's like those C++ questions that seek to test expertise by asking for answers to every esoteric point the test writer ever learnt two years ago. Meantime the experienced candidate long ago learnt to concentrate on proven and safe techniques and hasn't used esoteric crap for years.

Must be a Manager
Monday, December 01, 2003

Bah...My experience has been that a large percentage of technical questions on interviews serve no other purpose than to stroke the ego of the person asking the questions.

"WHAT!! You don't know what pin 27 of the Centronics interface does?!!? And you call yourself a developer!?!"

Even if you get past their mental masturbation, be very wary if you have to work with people who pull stupid interview stunts. They typically have deeper issues.

Mark Hoffman
Monday, December 01, 2003

If you're asking for specific pinpoints of information ("What is a pointer?"), then you're asking the wrong questions. If your interview is only 10 minutes, you're not giving a very good interview.

Last time I hired programmers, the interviews were 45-60 minutes depending on how long it took them to answer some of the tougher questions.

Things you want to see: how they solve problems. How well they communicate their answers. What happens when they don't know the answer to a question.

Also, if the position is "taxi driver" or "courier", not knowing where the empire state building is could be a legitimate reason not to hire them. Everything is context. :-)

Tim Sullivan
Monday, December 01, 2003

Just because an interviewer asks what 7! is, doesn't mean you have to blurt out 355687428096000.  A good interviewer would accept something that evaluates to the same thing, just as a computer would.

The person who boasted of using the printf question reminds us to try getting partial credit, rather than trying to make up some answer.  If the interviewer's bad, the interview is based on chance no matter what.  If the person's good but accidentally asks a bad question, you can salvage the situation.

We on this forum have the advantage of knowing what an interviewer must do, and we need to critique them and get to the point of their questions.

Tayssir John Gabbour
Monday, December 01, 2003

I think all interviews should have a problem solving component and a work history component.

If candidates can think, and they have a track record in the language/tools as demonstrated by example code then they make the short list.

I prefer to put candidates on the front foot by asking them to demonstrate some code from a prior project and explain it rather than asking them about something that I learned in the last week thats still fresh in my mind.

Realist
Monday, December 01, 2003

And you know that I neglected to type in a '1'. ;)

Tayssir John Gabbour
Monday, December 01, 2003

Also, I like to chat for 5 minutes in general about the weather, sports, whatever to try and get a 'feel' for the person. Interviews can be stressful for some people, in my experience putting people at ease is very worthwhile. It gets the best out of them.

Realist
Monday, December 01, 2003

"WHAT!! You don't know what pin 27 of the Centronics interface does?!!? And you call yourself a developer!?!"

Mark, pin 27 doesn't do anything, it's just ground. Since several other pins are ground as well, you can just disconnect it and it won't matter.

pinhead
Monday, December 01, 2003

A programmer's understanding of basic principles must be tested, How does a stack work? for instance. The principles are much more important than the language. A good or above average programmer will understand the principles and not be bothered by the language. A language can be learned quite fast, of course wih limitations, but trying to teach people basic programming principles ... not done.

Anycase the best question I ever got asked in an interview was right at the end: Who is your favourite superhero?

Of course Batman.

But hey, it might just be me ...

Daan Jacobs
Monday, December 01, 2003

We have a simple little C test we give to everyone. Now, name six different string comparison functions and the parameters, return values and parameter bounds for each.


Monday, December 01, 2003

"The person who boasted of using the printf question reminds us to try getting partial credit, rather than trying to make up some answer."

The question "will [x] compile?" doesn't really invite a partial credit answer, does it? Inviting partial credit would be to write it on a whiteboard and say "what do you think this would do?"

The reason I have a burn on this issue - I recently interviewed for a company. One interviewer asked "what is the default transaction isolation level for SQL Server?" I didn't know the semantics (though I know the effects). "Read Committed" he said, smugly.

Now interested, I asked "you know, I've only worked on small-scale apps, so that's never really been an issue for me. Have you had to deal with it in production?"

He said "no, I've never run into an issue with it." He grinned "but it makes a good interview question"

Uh, okay, but may I point out that something you've never used in the real world is NOT A GOOD INTERVIEW QUESTION. It's a *trivia* question. I'm not saying that SQL Server's transaction isolation levels are trivia, rather commenting on the attitude betrayed by the answer. (I'd be more impressed with "No, I've never run into it in real life. Hm. Maybe I should rethink asking that...")

One method I want to use more in the future is "what's wrong with this code?"

Philo

Wisdom trumps knowledge
Monday, December 01, 2003

>And I wouldn't work at a company that asked me what
>2+2 was in the interview.

We all know it's 5. But only for extremely large values of 2.

Patrik
Monday, December 01, 2003

Since several other pins are ground as well, you can just disconnect it and it won't matter...

Unless you're using flat (ribbon) cable.

I discovered what crosstalk was by doing that.

AJS
Monday, December 01, 2003

Specific stupid questions asked of me:

"Are you married?" and "Tell me something negative about yourself."

Gary (remove oink, 3)
Monday, December 01, 2003

>                                                          The principles are much
> more important than the language. A good or above average
> programmer will understand the principles and not be bothered by
> the language. A language can be learned quite fast, of course wih
> limitations, but trying to teach people basic programming principles
> ... not done.

This is a general truth that not enough HR departments/managers understand. I have a friend that is a whizz with older languages and is having problems getting a job, but his understanding of algorithms and structures is amazing.

whattimeisiteccles
Monday, December 01, 2003

And why should a programmer have to know every nit-picking syntax detail when the compiler will tell you if it's legal in a few seconds, or the IDE will tell you immediately as you type in a line?

NoName
Monday, December 01, 2003

Would you rather have someone who is just smart; or someone who is smart AND knows the language inside-out?

Mr Jack
Monday, December 01, 2003

If you find someone who is smart AND has the syntax details memorized, sure go ahead and hire that person.

However, if you are still interviewing people, it means you haven't yet found that perfect person.  By eliminating people instantly when they miss something in the syntax, you increase the possibility of hiring someone who memorized the syntax but isn't very good with the concepts.  I would rather hire someone with good concepts and not-so-good syntax than vice-versa.

NoName
Monday, December 01, 2003

> serve no other purpose than to stroke the ego of the person asking the questions

Yes.

On one of my last interviews, I was asked a highly technical question and I could tell that the interviewer was just dying to answer it himself. So I said something that sounded reasonable, and let him answer the rest of it the way he wanted to. He was very impressed with himself, and by extension, with me ;)

Portabella
Monday, December 01, 2003

> "Are you married?"
That's stupid, I agree.

> "Tell me something negative about yourself."
That one isn't stupid.

It's important to gauge the personality of the interviewee, and not just their skills.

You wouldn't want to hire someone who didn't know their own weaknesses, or someone insecure enough that they can't admit to any flaws, now would you?

Joe
http://www.joegrossberg.com

Joe Grossberg
Monday, December 01, 2003

Joe, I'm going to go out on a limb here & guess that you haven't done very much interviewing in your life.

Ask someone who spends an inordinate amount of time interviewing people, and you will discover that about 99% of the people give a canned response to canned interview questions.

Come on, have we not all heard that you are supposed to take the 'weakness' question and answer it with a 'strength' response?  i.e. "Since I demand such a high degree of productivity from myself, I am often disappointed in coworkers who don't share my strong work ethic".

tech recruiter
Monday, December 01, 2003

Philo, I don't know if the printf questioner was bad; I know I tend to exaggerate events from the night before.  I think it might be a good question that can be bad if you try hard to trick someone.  (If you talk about a bunch of green things and then suddenly ask what color means stop on a stoplight, the other person might say green.)

We do have to ask though, if the people he hired were successful on the new job.  It's not the only consideration, but it does count for something.

Actually, that printf question is one of the reasons why I don't currently care much for C.  Its compilers do all sorts of weird static checking, yet it's not typesafe.  I understand why that might be important for C's goals, but it's kind of scary to push C as a mainstream language.  It's a language for machine experts.

So the question is, should you be more stringent when interviewing C programmers, treating them like knifethrowers?

Tayssir John Gabbour
Monday, December 01, 2003

Um, 7! is 5040, not 355687428096000.  Good thing nobody asked you that on an interview.  :)

Phillip J. Eby
Monday, December 01, 2003

The New Yorker might have said "don't know" because he thought the interviewer was asking for the cross-street (or streets in this case), not the Avenue.

Pat Galea
Monday, December 01, 2003

Phillip, look at my correction one post down from the one you read.  I meant to say 17!, but somehow left off the 1.  Sigh, I don't typo english words on discussion forums, just the most embarrassing possible parts...

Tayssir John Gabbour
Monday, December 01, 2003

>>
Uh, okay, but may I point out that something you've never used in the real world is NOT A GOOD INTERVIEW QUESTION. It's a *trivia* question.
<<

That might apply for the question you were asked but the printf question is something that does pop up in real world C/C++ apps.  I've seen a handful of bugs in production code that resulted from lack of caution with the printf family of functions.  With the static type checking that goes on in C/C++, people tend to start relying on the compiler to catch things like that but with the printf family, you can't do that. 

It's reasonable to expect someone claiming C/C++ experience to be aware of some of the dangerous areas in the same manner that you'd expect someone with cooking experience to know that pans get hot.  If someone claiming cooking experience walks into a kitchen and immediately gets burned on something hot, you might begin to wonder how much experience they really have. 

SomeBody
Monday, December 01, 2003

I am just being picky: Empire State building is on the 34th Street, between the 5th and 6th Avenues.

In case anyone needs it for the interview…

coresi
Monday, December 01, 2003

"How are C++ objects laid out in memory?"

It's debatable; for some jobs I can see asking this question.  But for god's sakes, it's completely unspecified in the language specification.

Jason McCullough
Monday, December 01, 2003

> Empire State building is on the 34th Street

And 33rd! Big, innit? :-)

Pat Galea
Monday, December 01, 2003

I wonder if http://www.usabilitymustdie.com/jos/WW_All_Members.html is going to pick up the new Philo.

Just me (Sir to you)
Monday, December 01, 2003

Coresi & Pat - no hires for both of you. The address of the Empire State Building is on Fifth Avenue. ;-)

Philo

Wisdom trumps knowledge
Monday, December 01, 2003

Just me - I'm going to keep dropping behind anyway. New job is restricting my Joelizing.

Philo

Wisdom trumps knowledge
Monday, December 01, 2003

AJS,

You're right in general, but in this case the pins surrounding 27 pin 26 and 28, are also  ground.

If pin x-1 and x+1 were some high data rate signal and it was a flat ribbon cable as you say, which is the ideal situation for capacitive cross-coupling, then you certainly would be ill-advised to remove a ground on pin x.

pinhead
Monday, December 01, 2003

"> "Are you married?"
That's stupid, I agree."

It's illegal too (in the U.S.) You have to ask something like 'Is there any reason you wouldn't be able to travel/work late etc..'

Just testing
Monday, December 01, 2003

pinhead, yes 27, 28 & 29 may all be grounds, and be sitting next to each other on obe side of the connector, but once you go to ribbon cable they end up being interlaced with whatever is on the other side.

So the ribbon cable ends up being wired something like 13 / 27 / 14 / 28 / 15 / 29 / etc.

Of course, those who have never made cables will have no idea what this all means...

AJS
Tuesday, December 02, 2003

Philo,

If this is you who or somebody you care about who is being refused a job, be glad. Would you want to work for a company where the general attitude is like that?

I remember one interview where I had to list (it was on paper, with several, like 10 other candidates at the same time) the compiler settings for compiling a device driver (in the MS-DOS ages). Amongst a lot of other stupid questions. So as soon as I saw the questions, I quit. Never heared from them again. One month after that I had three other job offers.

Mark

Mark Tetrode
Tuesday, December 02, 2003

"I'm going to keep dropping behind anyway. New job is restricting my Joelizing."

It's about shame, Philo, not credit. ;-)

Just me (Sir to you)
Tuesday, December 02, 2003

Whoa! Thanks AJS, learned something new!

(Personally I love cables, especially the pretty ones with funky colors.)

pinhead
Tuesday, December 02, 2003

"Joe, I'm going to go out on a limb here & guess that you haven't done very much interviewing in your life."

Heh ... no I haven't. I've had four tech interviews and two tech jobs in my life.

I did, however, get asked that question when I was in college (my first interview, for a summer internship), and it through me for a loop.

And, yes, when I asked, in the university's career services seminar, she proposed the exact same "canned" response you cite: say you're a perfectionist or the like.

Thx for the insight,
Joe
http://www.joegrossberg.com

Joe Grossberg
Wednesday, December 03, 2003

Ack ... "threw", not "through".

Joe
http://www.joegrossberg.com

Joe Grossberg
Wednesday, December 03, 2003

"> "Are you married?"
That's stupid, I agree."

It's illegal too (in the U.S.) You have to ask something like 'Is there any reason you wouldn't be able to travel/work late etc..'

I don't know about that--I think the interviewer would be left confused.  After all, wouldn't a married person answer NO ('Are you saying I get to leave on occassion, and I can say it's job-related?!')  Believe you me, it is only a single person who would feel they can't leave their sweetheart behind--after all, (s)he may end up finding another in the interviewee's absence.

Patrick Boland
Wednesday, August 18, 2004

*  Recent Topics

*  Fog Creek Home