Interview questions?
-
Hi, I'm giving an interview (2nd round) to a potential VC/C/C++ software engineer (4 years experience) later this afternoon. What sort of questions should I ask? Do you have some good sample technical questions you can share? I'd kind of like to test this guy's technical knowledge and see how he handles both easy and difficult questions... Thanks! Chris
Don't ask those stupid riddles and "write the algorithm" questions. (At one group interview years ago, one guy asked me to show how I would reverse a string. I wrote strrev(pStr); on the board and he got really ticked.) I've seen no evidence that there are certain questions you can ask that will tell you if this engineer is the right one. I know engineers than can explain to you, in detail, the peformance of a btree versus a trie, but can't write practical code that works. How about this: Explain to me a recent serious programming problem you encountered and how you solved it. When you go over his resume, look at the products he has worked on. Do a little research. Find out what the downfalls of these products are and ask questions pertaining specifically to that; As in, "How did you solve the problem of ....?" The point here isn't that there are right answers. I strongly suggest asking a few questions about how familiar the developer is with the tools, especially the debugger. Don't ask trick questions--I'm a very good debugger but I don't remember what key strokes do what, I just do it. (My daughter asked me the other day to teach her how to tie a tie; I do it so automatically, that I found it very difficult to explain it.) Afterward, call up his references and ask them pointed questions using the interview as a guideline (that is, if you are in a state, like Utah, that idemnifies you and the references from an court action. If not DON'T do this this step, you'll have to trust your instincts as to whether the person is on the level.) Of course, you need to be clear in you own mind what you want. Do you want a yes man? Or do you want someone who questions your proposals [with valid reasons]? (I suggest the latter but many "managers" can't handle it, certainly many of mine couldn't.) Oh yeah, and ask him/her what developer books and/or articles they've read recently. Ask him/her from what web site they "borrow" code. (Okay, this last is a trick question; anyone who assures me they always write their own code for everything won't be hired by me.)
-
Tim Smith wrote: There are many people, including myself who feel that a math background goes a long way towards being a good programmer. You get analytical thinking, problem solving, and logic. I failed math at school, failed it utterly and miserably. Failed it so badly my future son can feel my math teachers pain. Yet I consider my problem solving to be rather good. I am logical (no really I am logical when it comes to doing programming, the rest of me is not, as you have seen :-D) and being partly a systems analyst consider my analytical thinking to be good. I don't think it has much to do with math, just clear, logical, good thinking, which are helpful attributes if you want to do well at math (I assume.) Though as someone awhile back pointed out I probably could have done well at math, I just honestly did not care about school math. It was pointless to me. Also what do you consider to be "good at math" to mean? Algebra? Geometry? Adding? Are all math skills condusive to good programming? It seems to me more like the type of person who is good at math (logical, analytical etc.) is suited for programming, not because they can do complex algebraic equations (which I can't because why is the x important? and what does the y represent? and will this make the world better?) but because of those traits you mentioned. Is math maybe your way of developing those traits?
Paul Watson
Bluegrass
Cape Town, South AfricaWell, actually you hit on what I really wanted to say. It isn't the "can you compute the area of X" math, but the analytical part of math. What would be years 3-4 for someone getting their math degree. Calculus doesn't help me to program. But the logical training I got trying to prove math from the ground up in analysis class was invaluable. Tim Smith "Programmers are always surrounded by complexity; we can not avoid it... If our basic tool, the language in which we design and code our programs, is also complicated, the language itself becomes part of the problem rather that part of the solution." Hoare - 1980 ACM Turing Award Lecture
-
Hi, I'm giving an interview (2nd round) to a potential VC/C/C++ software engineer (4 years experience) later this afternoon. What sort of questions should I ask? Do you have some good sample technical questions you can share? I'd kind of like to test this guy's technical knowledge and see how he handles both easy and difficult questions... Thanks! Chris
Besides tests previously spotted on former posts, we give potential programmers some code snippets (4 to 6) in a piece of paper and ask them to analyze the code. Written analysis. Kind of a "peer review" task. The code have tons of good/bad programming habits and he has to spot them. We usually use this test to assess: 1. How well does he write? He'll write documentation this way! 2. What's his maturity level (as a programmer)? 3. What are his strengths: optimize, correct or organize code? 4. How well he'd mantain someone else's code? Concussus surgo. When struck I rise.
-
Hi, I'm giving an interview (2nd round) to a potential VC/C/C++ software engineer (4 years experience) later this afternoon. What sort of questions should I ask? Do you have some good sample technical questions you can share? I'd kind of like to test this guy's technical knowledge and see how he handles both easy and difficult questions... Thanks! Chris
Here are a few questions I use in interviews: What has been the most satisfying experience at work ? The most frustrating ? The most obscure problem you have encountered ? Did you solve it and how ? And a VC++ question: What bugs/limitations have you found in MFC ? Hope this is of help. Elaine (fluffy tigress emoticon) Would you like to meet my teddy bear ?
-
My favorite one is: calculate the sum of all integers from 1 to n. If guy writes a loop for that, you know his math background is lacking ;) Tomasz Sowinski -- http://www.shooltz.com
To some its a six-pack, to me it's a support group
-
Tomasz Sowinski wrote: you know his math background is lacking Just curious but what on earth has math got to do with everyday programming?
Paul Watson
Bluegrass
Cape Town, South AfricaPaul Watson wrote: Just curious but what on earth has math got to do with everyday programming? I've often wondered about this. My math's skills are very ropey (partly down to bad teaching - I sort of got left behind because my geometry drawing was bad and from then on I was kind of ignored). I wonder if I would be a better programmer if I was better/enjoyed maths. I've been programming for a long time now and never really felt that maths was essential. Maths has very little relevence to my work with MIS and telephony systems. If I need a specialised routine, thats what libraries are for. Michael Programming is great. First they pay you to introduce bugs into software. Then they pay you to remove them again.
-
This question has only one point: you may check if programmer starts coding right away or spends 30 second thinking about the solution. Very important aspect, IMHO. Tomasz Sowinski -- http://www.shooltz.com
To some its a six-pack, to me it's a support group
Tomasz Sowinski wrote: This question has only one point: you may check if programmer starts coding right away or spends 30 second thinking about the solution. Very important aspect, IMHO. Yes, I agree that it's important to hire someone that can get started right away. But why not do it with a practical question? As I implied, unless the candidate has had recent exposure to "cool" math algorithms, he just isn't going to know the answer to this one. (I didn't know it before and heck, I've already forgotten it. :-)) But something like, "How do you make a copy of an array?", is IMHO a better way to measure a candidate's readiness for the job. And by the way, spending an extra 30 seconds (or minutes) thinking about how to solve a problem is usually OK if it yields maintainable code that does the job well, IMHO. Plus, if you're getting payed by the hour, even better ;). Regards, Alvaro Insanity: doing the same thing over and over again and expecting different results. - Albert Einstein
-
Really you're not allowed to ask for code? I've been to interviews before and had to write some, admittedly it was VB (no flame follow-ups please :) but I still wrote it. What's the reasoning behind it? Some kind of copyright thing? What about asking for flow-chart type stuff? -- Paul "I need the secure packaging of Jockeys. My boys need a house!" - Kramer, in "The Chinese Woman" episode of Seinfeld MS Messenger: paul@oobaloo.co.uk Sonork: 100.22446
Yes - It is surprising, but it turns out that unless your tests can be defended in court, avoid it altogether. For more detail : read up on (Supreme Court Griggs Vs Duke Power) http://www.employeeselect.com/testLegal.htm[^] Some more rules (if u wanna stay out of trouble) 1. Avoid problem solving questions that are not standardized.(not asked to every potential candidate ) (eg. "What is the command in CVS to create a branch ?" to one candidate and "What is CVS ?" to another one). This is a clear example of aribitrary testing. The employee has the right to sue the company under these circumstances. 2. No role-playing questions (eg. Sell me that chair) 3. No questions at all about personal stuff, hobbies, lifestyle, general attitudes. 4. Situational questions : (What would u do if you were at a customer site and the product refused to print ) Of course,most interviewees wont bother, but only takes one pissed off employee to really get ur company in a legal mess :omg: I am frankly surprised so many in CP are encouraging tests in interviews. Any comments ? I would like to hear your views.. Vivek
-
This question has only one point: you may check if programmer starts coding right away or spends 30 second thinking about the solution. Very important aspect, IMHO. Tomasz Sowinski -- http://www.shooltz.com
To some its a six-pack, to me it's a support group
Just for the record, I think it is an excellent question !! It took me more than 30secs to realise the solution though :-( as it appeared way too easy. Regardz Colin J Davies
Sonork ID 100.9197:Colin
You are the intrepid one, always willing to leap into the fray! A serious character flaw, I might add, but entertaining. Said by Roger Wright about me.
-
Hi, I'm giving an interview (2nd round) to a potential VC/C/C++ software engineer (4 years experience) later this afternoon. What sort of questions should I ask? Do you have some good sample technical questions you can share? I'd kind of like to test this guy's technical knowledge and see how he handles both easy and difficult questions... Thanks! Chris
Aside from the technical questions (which I often find misleading, I much prefer to get him to write some actual code). I find the best way to judge a person's technical knowledge is to get him to talk about the projects he has worked on in the past. Find out what he's done and ask about the reasons he did things the way he did. Michael Programming is great. First they pay you to introduce bugs into software. Then they pay you to remove them again.
-
Maximilien wrote: Remember, your job is to hire someone, not break him! Agreed. Also, if you are going to be working with the guy, the most important thing is to decide if you and your co-workers will like him. Andy Cowenhoven
Andy Cowenhoven wrote: Also, if you are going to be working with the guy, the most important thing is to decide if you and your co-workers will like him. This is asking for trouble in the US. You are not allowed to test a candidate for *general likeability* (Griggs vs Duke Power) - see my other post below. Yes, as an unspoken rule - you may only select candidates u feel comforatble with. The candidate can sue the company big-time if you explicity test him or her for likeability or other non job-related criteria. Therefore, avoid all questions that may be construed to be non-job related, such as: 1) Do you like beer ? we have a beer party every Friday. 2) I hope u like to play foosball, we have a league in the office. etc,etc u get the idea Vivek
-
Tomasz Sowinski wrote: This question has only one point: you may check if programmer starts coding right away or spends 30 second thinking about the solution. Very important aspect, IMHO. Yes, I agree that it's important to hire someone that can get started right away. But why not do it with a practical question? As I implied, unless the candidate has had recent exposure to "cool" math algorithms, he just isn't going to know the answer to this one. (I didn't know it before and heck, I've already forgotten it. :-)) But something like, "How do you make a copy of an array?", is IMHO a better way to measure a candidate's readiness for the job. And by the way, spending an extra 30 seconds (or minutes) thinking about how to solve a problem is usually OK if it yields maintainable code that does the job well, IMHO. Plus, if you're getting payed by the hour, even better ;). Regards, Alvaro Insanity: doing the same thing over and over again and expecting different results. - Albert Einstein
Being able to solve the problem as Tomasz suggested with an algorithm shows how the candidates mind works. I don't think he is suggesting that is the only factor for determining who get's the job. When I first started programming, 20 ++ yrs ago we always searched for algorithms before trying structured methods. There were more advantages then, but one main advantage remains, well written algorithms don't fail like most structures do thus less bugs. Regardz Colin J Davies
Sonork ID 100.9197:Colin
You are the intrepid one, always willing to leap into the fray! A serious character flaw, I might add, but entertaining. Said by Roger Wright about me.
-
Aside from the technical questions (which I often find misleading, I much prefer to get him to write some actual code). I find the best way to judge a person's technical knowledge is to get him to talk about the projects he has worked on in the past. Find out what he's done and ask about the reasons he did things the way he did. Michael Programming is great. First they pay you to introduce bugs into software. Then they pay you to remove them again.
Michael P Butler wrote: (which I often find misleading, I much prefer to get him to write some actual code). It would be ok , if the test were standardized or at least had a consistent level of difficulty. Isnt this arbitrary ? What if you asked one person to write a "hello world" and another person to write a "recursive loop" ? This may happen becuase there may be different interviewers. - V
-
Tomasz Sowinski wrote: This question has only one point: you may check if programmer starts coding right away or spends 30 second thinking about the solution. Very important aspect, IMHO. Yes, I agree that it's important to hire someone that can get started right away. But why not do it with a practical question? As I implied, unless the candidate has had recent exposure to "cool" math algorithms, he just isn't going to know the answer to this one. (I didn't know it before and heck, I've already forgotten it. :-)) But something like, "How do you make a copy of an array?", is IMHO a better way to measure a candidate's readiness for the job. And by the way, spending an extra 30 seconds (or minutes) thinking about how to solve a problem is usually OK if it yields maintainable code that does the job well, IMHO. Plus, if you're getting payed by the hour, even better ;). Regards, Alvaro Insanity: doing the same thing over and over again and expecting different results. - Albert Einstein
Alvaro Mendez wrote: As I implied, unless the candidate has had recent exposure to "cool" math algorithms, he just isn't going to know the answer to this one First of all, the formula for sum(1..n) is hardly 'cool' - it's firetrucking basic one. Secondly, you don't need to remember the formula, because you can easily derive it yourself. Just notice that you can add 1 to n, 2 to (n-1), 3 to (n-2) etc... Tomasz Sowinski -- http://www.shooltz.com
To some its a six-pack, to me it's a support group
-
Tomasz Sowinski wrote: Surely it has nothing to do with javascript Hardy har har. Tomasz Sowinski wrote: I'm not sure what do you mean by 'everyday programming' The kind of programming that creates apps that users use to get their job done better. I fully agree math is important if you are working for Intel or doing scientific applications, but how many of us actually do that kind of programming? Most of us are data-managers. We make systems which allow data input, data output, data processing and data transformation. Not much else. So far I have had very little need for math beyond adding. Where there are calculations to do, they are hardly complex and are understandable because instead of
x2 * 2y / cos(50)we have((warrantPrice * totalWarrants) / strikePrice) - (sharePrice2)or something. All in all though the calculations are (while critical) a very small part of the code you write.Paul Watson
Bluegrass
Cape Town, South AfricaPaul Watson wrote: All in all though the calculations are (while critical) a very small part of the code you write Hey Something we agree on!!!! We did a survey ( line count type ) of our apps which are mainly insurance type applications ( calculating rates etc.. ) and the actual percentage of calulation code vs UI and printing was 7% thats all 7 lines out of 100. And a lot of that was stuff to calculate the strange and murkey underwriting logic - not so much number crunching. Richard We are called the nation of inventors. And we are. We could still claim that title and wear its loftiest honors if we had stopped with the first thing we ever invented, which was human liberty. Mark Twain- Foreign Critics speech, 1890
-
Tomasz Sowinski wrote: Surely it has nothing to do with javascript Hardy har har. Tomasz Sowinski wrote: I'm not sure what do you mean by 'everyday programming' The kind of programming that creates apps that users use to get their job done better. I fully agree math is important if you are working for Intel or doing scientific applications, but how many of us actually do that kind of programming? Most of us are data-managers. We make systems which allow data input, data output, data processing and data transformation. Not much else. So far I have had very little need for math beyond adding. Where there are calculations to do, they are hardly complex and are understandable because instead of
x2 * 2y / cos(50)we have((warrantPrice * totalWarrants) / strikePrice) - (sharePrice2)or something. All in all though the calculations are (while critical) a very small part of the code you write.Paul Watson
Bluegrass
Cape Town, South AfricaPaul Watson wrote: but how many of us actually do that kind of programming? me! Paul Watson wrote: Most of us are data-managers. i absolutely f'in hate data pushing. Paul Watson wrote: All in all though the calculations are (while critical) a very small part of the code you write. yeah, in most places that's probably true. but not in my house! :) -c
"Kate said / The flowers of intolerance and hatred / Are blooimg kind of early this year / Someone's been watering them. -- Robyn Hitchcock, Devil's Radio
-
Maximilien wrote: second round ? does that mean he was already interviewed ? if so I would assume that all the technical questions were already asked and answered, and that he answered correctly ! Actually it could be quite the opposite, many times companies have a front ( I like to call them HR departments) where they have associates try to run you through the ringer with some techno-babble (that of which they don't understand) and if they like you, you progress to stage two; speaking with someone who is in the IT department :cool:. My suggestion is keep in mind that there are always many ways to do the similar task. Given Tomasz question at the top, he may initially list off a for loop where you use a simple math algorithm. Then again you probably don't know everything and at sometime he may actually shed some light as to a more efficient way to do another process than you. Don't initially strike them for good because they didn't answer a question exactly as you thought was *best*. Nick Parker
Nick Parker wrote: ( I like to call them HR departments) IMHO: I think they are scum, and often cause the wrong people to see the IT departments. Regardz Colin J Davies
Sonork ID 100.9197:Colin
You are the intrepid one, always willing to leap into the fray! A serious character flaw, I might add, but entertaining. Said by Roger Wright about me.
-
Being able to solve the problem as Tomasz suggested with an algorithm shows how the candidates mind works. I don't think he is suggesting that is the only factor for determining who get's the job. When I first started programming, 20 ++ yrs ago we always searched for algorithms before trying structured methods. There were more advantages then, but one main advantage remains, well written algorithms don't fail like most structures do thus less bugs. Regardz Colin J Davies
Sonork ID 100.9197:Colin
You are the intrepid one, always willing to leap into the fray! A serious character flaw, I might add, but entertaining. Said by Roger Wright about me.
Colin^Davies wrote: Being able to solve the problem as Tomasz suggested with an algorithm shows how the candidates mind works. Fair enough, but what ends up happening is one of these things: 1. The candidate knows the answer cold, either because his name is Tomasz or because he read (and memorized) Tomasz's response above. :-) 2. The candidate spends a few seconds (or minutes) trying to figure out the algorithm. So much for that theory of him being able to jump in and start coding right away. :-) 3. The candidate gets nervous and comes up with a not-so-cool solution, like a loop or recursive function. 4. The candidate is a dope and gives an incorrect or stupid answer. Insanity: doing the same thing over and over again and expecting different results. - Albert Einstein
-
i know some good programmers who use the mouse for copy/paste. i don't understand it, at all. i have the same feeling about them that i do about people who enable every toolbar an app offers, leaving them all of two inches of actual workspace. then there are the people who enable/disable breakpoints or start builds, or any number of things from the toolbar (just use the keyboard!) it infuriates me. (have i insulted everyone here yet? :) ) -c
"Kate said / The flowers of intolerance and hatred / Are blooimg kind of early this year / Someone's been watering them. -- Robyn Hitchcock, Devil's Radio
I have two modes: 1. Keyboard - used when editing code 2. Mouse - used when debugging code So I basically use the mouse to click on the buttons for stepping through the code and setting breakpoints, etc. But if I'm just writing code, I make full use of the keyboard shortcuts -- Ctrl+S being my favorite one. :-) Regards, Alvaro Insanity: doing the same thing over and over again and expecting different results. - Albert Einstein
-
i know some good programmers who use the mouse for copy/paste. i don't understand it, at all. i have the same feeling about them that i do about people who enable every toolbar an app offers, leaving them all of two inches of actual workspace. then there are the people who enable/disable breakpoints or start builds, or any number of things from the toolbar (just use the keyboard!) it infuriates me. (have i insulted everyone here yet? :) ) -c
"Kate said / The flowers of intolerance and hatred / Are blooimg kind of early this year / Someone's been watering them. -- Robyn Hitchcock, Devil's Radio
I once saw a programmer (yes, a programmer) do the following: Using the mouse: 1. click on a word 2. highlight the word 3. pull down the Edit menu 4. click Cut 5. click on the new position 6. pull down the Edit menu 7. click Paste Guess how far he needed to move the word. Marc