Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie
Java technical test
Options
-
17-03-2014 7:08pm#1
Comments
-
-
-
-
-
-
Advertisement
-
I can't really give you specifics because they were aimed at my own experience how I had used sql to solve problems in work, creating tables, inserting etc.Can you please post some of the whiteboard and SQL questions if you remember them
The Design session was also based around my work I had made an app and they asked me to show a design on how you interacted with it how I would improve it now, how I would scale it up.
That's all I can remember the whole thing was a blur but they were lovely and easy to deal with.0 -
-
-
Ye they are crap but I suppose every decent developer should know about the ternary operator:
max = (a > b) ? a : b;
but after reading this question I googled about a bit and found that the Java math library has a max() that does this for you:
Math.max(10,20); returns 20
So help me, if the crowd I was interviewing for accepted either of those answers, I'd have to think long and hard about working there.
What exactly do you think the ternary operator and the max() functions are under the hood but if statements? So either they just mean don't use the if keyword in which case, that's a really horrible test and indicates a level of competence that hints to long nights fixing stupid mistakes; or they really mean don't use comparisons at all, which is a bitch of a question.
First way I can think of if it's the latter demonstrates how ugly a question it is (I'll do it for chars here in C so it's more obvious):#include <stdio.h> unsigned char max(unsigned char a, unsigned char b) { unsigned char answer[65536]; int answerIndex = 0; for (int i = 0; i<=255; i++) { for (int j = 0; j<=255; j++) { /* this is setup, so using if here isn't * cheating because we could just enter * all this by hand... ewww */ answer[i<<8|j] = i > j ? i : j; } } /* here's the real max function */ answerIndex = a << 8 | b; return answer[answerIndex]; /* good programmers should now feel the need for a shower */ } int main () { printf ("max (128,255) = %i\n", max(128,255)); return(0); }
Nasty. In my defence, it's late and I'm tired
If I was running a tech interview and someone came up with that answer to that question but said it was messy, I'd pass them because interviews are stressful, but I'd send it back if it showed up in a code review in day-to-day work as being too mucky.
There is a much, much cleaner way if you bittwiddle, because signed ints will be in two's complement format on pretty much every machine I can think of so you can check the sign without doing comparisons something like this:int diff = a - b; int complement = (diff >> 31) & 0x01; int max = a - k * diff;
So you're grabbing the two's complement bit of diff with the shifting there (we're assuming 32 bit ints, but that's a reasonably safe bet in modern C/C++) (and really the masking shouldn't be needed, I just have really old habits ), and it'll be either 0 or 1 so k*diff will be 0 if a > b and it'll be equal to a - b if a < b; therefore max is either just a (if a > b) or a - (a - b) which is just b, if a < b.
Here's the thing. That's a nice question if you want someone who can do low-level bit-twiddling in C or C++, but the whole point of languages like Java was to get the hell away from that kind of thing. So it's still a stupid Java question. You might as well ask someone to write an associative array in python from scratch instead of using dictionaries.0 -
Moderators, Science, Health & Environment Moderators, Social & Fun Moderators, Society & Culture Moderators Posts: 60,092 Mod ✭✭✭✭
Join Date:Posts: 59268
I went for an android interview, 3rd and final round. The whole test was in C/++ and was got to do with "do this without using inbuilt functions and bonus marks for the more shifting of bits you do". At no stage in the interviews was anything java or android really asked about, an odd question here and there. It was disgusting, and they were offering 20k.
The whole point is that I'm abstracted from that stuff, waste of time, and I had no idea how to do anything with bit shifting anyway. It would take me a good while to even follow the logic in Sparks post 0 -
Advertisement
-
-
That's cheatingThere are other approaches though, stick them into a sortable collection, sort and then pick the first (or last) element depending on the sort order for example. The non-cheating answer to the harder version of the question kindof means that if you checked the assembly, there'd be no conditionals in there. C'mon, play the game, it's fun
That's why it's such an awful question for a language like Java (but actually, kindof a good question if bittwiddling is going to be a thing in the job).I'd use the question to see if someone can come up with imaginative solutions that others can understand, not low-level crap that the next guy is going to have to google to understand before he can attempt to make a change.
And if you've been coding for a while now, you know all these things aren't wierd or exotic, they were commonplace a decade or two ago, and still are in some areas of coding at low level or in places where really high performance is a factor. There are whole lists of them (this one looks new but it's just been reformatted from the version I first saw in 1993, and it was very old then; and there's also hakmemc and the famous Graphics Gems collections) and people used to read those lists quite a lot.
So for low-level stuff, this is actually okay (but if the applicant is stressed, it could go horribly wrong) - the problem is that it's a horrible high-level question and shouldn't be on a java test at all.0 -
No it wouldn't, if you weren't at work and weren't distracted, you wouldn't even get through a cup of coffee in the time it took to grok it.Tar.Aldarion wrote: »It would take me a good while to even follow the logic in Sparks post
You want something hard to follow at that level, go look up John Caramack's inverse square estimation routine in the original Quake engine code:float InvSqrt(float x){ float xhalf = 0.5f * x; int i = *(int*)&x; // store floating-point bits in integer i = 0x5f3759d5 - (i >> 1); // initial guess for Newton's method x = *(float*)&i; // convert new bits into float x = x*(1.5f - xhalf*x*x); // One round of Newton's method return x; }
There's an entire academic paper written about that one
(Well, okay, the paper is mainly about the choice of 0x5f3759d5 as the magic number in there, but it's still shiny).0 -
Hmm I'm guessing they would frown on it but below is a really simple way to do it in java.So help me, if the crowd I was interviewing for accepted either of those answers, I'd have to think long and hard about working there.
What exactly do you think the ternary operator and the max() functions are under the hood but if statements? So either they just mean don't use the if keyword in which case, that's a really horrible test and indicates a level of competence that hints to long nights fixing stupid mistakes; or they really mean don't use comparisons at all, which is a bitch of a question.
First way I can think of if it's the latter demonstrates how ugly a question it is (I'll do it for chars here in C so it's more obvious):#include <stdio.h> unsigned char max(unsigned char a, unsigned char b) { unsigned char answer[65536]; int answerIndex = 0; for (int i = 0; i<=255; i++) { for (int j = 0; j<=255; j++) { /* this is setup, so using if here isn't * cheating because we could just enter * all this by hand... ewww */ answer[i<<8|j] = i > j ? i : j; } } /* here's the real max function */ answerIndex = a << 8 | b; return answer[answerIndex]; /* good programmers should now feel the need for a shower */ } int main () { printf ("max (128,255) = %i\n", max(128,255)); return(0); }Nasty. In my defence, it's late and I'm tired If I was running a tech interview and someone came up with that answer to that question but said it was messy, I'd pass them because interviews are stressful, but I'd send it back if it showed up in a code review in day-to-day work as being too mucky.
There is a much, much cleaner way if you bittwiddle, because signed ints will be in two's complement format on pretty much every machine I can think of so you can check the sign without doing comparisons something like this:int diff = a - b; int complement = (diff >> 31) & 0x01; int max = a - k * diff;
So you're grabbing the two's complement bit of diff with the shifting there (we're assuming 32 bit ints, but that's a reasonably safe bet in modern C/C++) (and really the masking shouldn't be needed, I just have really old habits ), and it'll be either 0 or 1 so k*diff will be 0 if a > b and it'll be equal to a - b if a < b; therefore max is either just a (if a > b) or a - (a - b) which is just b, if a < b.
Here's the thing. That's a nice question if you want someone who can do low-level bit-twiddling in C or C++, but the whole point of languages like Java was to get the hell away from that kind of thing. So it's still a stupid Java question. You might as well ask someone to write an associative array in python from scratch instead of using dictionaries.
Actually I'd say they'd ask for another way as well.public int getMaxWithNoIFWhile(int a, int b) { while(a>b) { return a; } return b; } public int getMaxWithNoIFFor(int a, int b) { for(int i;a>b;i++) { return a; } return b; }0 -
-
-
There's probably another bittwiddling way to do it, and you could do a version of the disgusting lookup table solution that's even worse (instead of using it as a lookup table, use it as a jump table where you goto the address in the lookup table and either land at a return(a) or return(b) statement), but even I won't code that
Thing is, there aren't that many ways to do stuff this low-down, which is yet another strike against this as an interview question...0 -
Moderators, Science, Health & Environment Moderators, Social & Fun Moderators, Society & Culture Moderators Posts: 60,092 Mod ✭✭✭✭ Join Date:Posts: 59268
Gonna have a read of that cheers, I love reading the work of FBs newest employees anywayNo it wouldn't, if you weren't at work and weren't distracted, you wouldn't even get through a cup of coffee in the time it took to grok it.
You want something hard to follow at that level, go look up John Caramack's inverse square estimation routine in the original Quake engine code:float InvSqrt(float x){ float xhalf = 0.5f * x; int i = *(int*)&x; // store floating-point bits in integer i = 0x5f3759d5 - (i >> 1); // initial guess for Newton's method x = *(float*)&i; // convert new bits into float x = x*(1.5f - xhalf*x*x); // One round of Newton's method return x; }
There's an entire academic paper written about that one
(Well, okay, the paper is mainly about the choice of 0x5f3759d5 as the magic number in there, but it's still shiny).
Distracted is right, all of our products now echo my name to the screen temporarily
hmm another one I was asked was convert string to int without using atoi etc, get cracking lads.0 -
Right, that's it, you're banned.Tar.Aldarion wrote: »FBs newest employee
shutupshutupshutupshutupshutupshutupshutupshutupSHUTUP!
Bug report: Tar is a bit narcissistic and now our product is enabling himDistracted is right, all of our products now echo my name to the screen temporarily
Hmmmmm..... ascii or truly evil unicode?hmm another one I was asked was convert string to int without using atoi etc, get cracking lads. 0 -
Tar.Aldarion wrote: »hmm another one I was asked was convert string to int without using atoi etc, get cracking lads.
In Python:digits = "0123456789" result = 0 for d in input_string: result = result * 10 + digits.index(d)
Adapt as necessary for less... elegant languages.
0 -
Advertisement
-
I love me some Python, but even I wouldn't have said that while Ruby was listeningoscarBravo wrote: »less... elegant languages. 0 -
Tar.Aldarion wrote: »I went for an android interview, 3rd and final round. The whole test was in C/++ and was got to do with "do this without using inbuilt functions and bonus marks for the more shifting of bits you do". At no stage in the interviews was anything java or android really asked about, an odd question here and there. It was disgusting, and they were offering 20k.
The whole point is that I'm abstracted from that stuff, waste of time, and I had no idea how to do anything with bit shifting anyway. It would take me a good while to even follow the logic in Sparks post
Third round, they gave you a C++ exam (for Android?!), and offering 20k?
Jaysus! That's an insult. 3 interviews deep, and an exam that doesn't really make sense.0 -
-
I have this theory that certain people grok certain languages and that the mapping is unique to each person. Me, I grok C and python and I know a few others well enough to do good work in them, but I keep looking at Ruby and thinking I'd like to learn it but it never sits still in my head long enough to do anything worthwhile with. Still strikes me a a really pretty language though.oscarBravo wrote: »I can't get to love Ruby. The syntax just leaves me cold. After almost three decades of professional software development, I think I've found my One True Love in Python.
But then, Erlang made lots of sense to me from day one and I'm told that's a bit abnormal, so I might be a bad data point 0 -
Moderators, Science, Health & Environment Moderators, Social & Fun Moderators, Society & Culture Moderators Posts: 60,092 Mod ✭✭✭✭ Join Date:Posts: 59268
Yeah, apparently was down to the last two, adn teh hour long phone interview that I did also...focused on RTOS. Wut.Third round, they gave you a C++ exam (for Android?!), and offering 20k?
Jaysus! That's an insult. 3 interviews deep, and an exam that doesn't really make sense.
Python is cheatingoscarBravo wrote: »In Python:digits = "0123456789" result = 0 for d in input_string: result = result * 10 + digits.index(d)
Adapt as necessary for less... elegant languages.
Python, what can i say, I never did it until i started working here, it is beautiful.oscarBravo wrote: »I can't get to love Ruby. The syntax just leaves me cold. After almost three decades of professional software development, I think I've found my One True Love in Python.0 -
-
-
-
-
Advertisement
-
Advertisement