Technical Test

les_steaks

Hi all, I currently looking to change jobs. I'm a VB develpoer with 5 years experience. Contacted a few agencies, one mentioned a senior programming role with Citigroup. Haven't done an interview or anything yet but he mentioned a 3 hour technical test. Anyone know what to expect here?
Cheers, Les

PhantomBeaker

Well, I'm assuming some of it will be down to making sure you know the language and general caveats of it and the like.

I had a short technical test with IBM for a summer internship, and it essentially boiled down to asking questions about pointers and the like. For example (this is in java though)

String s1 = "Hello world";
String s2 = "Hello world";

if(s1 == s2){
System.out.print(1);
} else{
System.out.print(2);
}

What is printed in this case.. that sort of stuff. Basically proving you can understand code fragments and the subtleties behind them.

Maybe asking how you invoke your programs, how you'd write libraries and how you'd reference them. Basically making sure you know your stuff.

You might also be asked questions about your technical knowledge... like questions about how the xml and xslt technologies relate, or similar stuff. We were also asked a rake-load of these questions for the internship.

grahamor

i work for citigroup, was meant to do a tech test and never did, thank god. i think PhantomBeaker is right though, just understanding coding in general and how things are done is what you need to know (sequence, selection, iteration)

GreeBo

PhantomBeaker wrote:

String s1 = "Hello world";
String s2 = "Hello world";

if(s1 == s2){
System.out.print(1);
} else{
System.out.print(2);
}

What is printed in this case..

"1"
What do I win?

Hobbes

GreeBo wrote:

"1"
What do I win?

bit more to it then that. How about.

String s1 = new String("Hello World");
String s2 = new String("Hello World");

(which I suspect what the actual question was)

GreeBo

Hobbes wrote:

bit more to it then that. How about.

String s1 = new String("Hello World");
String s2 = new String("Hello World");

(which I suspect what the actual question was)

ahh but thats a totally different question!

The original was harder because it sucked you into the object comparison v's string value comparison "trap", though I suspect it was a type

PhantomBeaker

Heh, what do you know... the answer to that question IS 1, the way I asked it.

And yes, I suppose I did mean the s1 = new String etc etc...

That's an interestink thing to know about the java compiler... it does optimise somewhat. :rolleyes:

...cool...

GreeBo

PhantomBeaker wrote:

Heh, what do you know... the answer to that question IS 1, the way I asked it.

And yes, I suppose I did mean the s1 = new String etc etc...

That's an interestink thing to know about the java compiler... it does optimise somewhat. :rolleyes:

...cool...

as I said, what do I win?

PhantomBeaker

GreeBo wrote:

as I said, what do I win?

My love and respect.

Of course by love I don't mean the love between a man and a woman, but more the love between a programmer and the task of having to debug object-oriented code in assembly (think "gcc -s").

And by respect, as much respect as is due to one who probably compiled it to see what it did.



Aoife

plenderj

I went for a job as a senior dev in London using VB primarily. I walked in, and they said "Right, you're a VB MVP. Do you mind if we do this interview in C#?".

The first things I had to do up on a whiteboard in front of them was write code to multiply two numbers together without using the multiply operator.
Then I had to code it recursively... urgh!

PhantomBeaker

You know, that's going into my repetoire of tests to see if I know a language or not... multiplication without the multiplication operator.

It's simple when you think about it outside an interview, but I'd probably brick it if that was given to me as part of one.

(Edit: ok, just implemented the various ways I was thinking about doing it... after some initial logic bug hunting, it's all good - Thanks for the question, it was mildly entertaining for a short while : ) )

plenderj

Now do it recursively

PhantomBeaker

Oh, well, I have about 4 functions. 2 are recursive and 2 are iterative. So basically two different ways of doing it, but both done iteratively and recursively.

Trust my black heart to think of the more difficult version first! When I thought of the difficult version I thought "Oooh, that's a bad one to ask in an interview" until I thought of the easy way and realised "Oh. The horrible part of the question is realising exactly what's being asked, not actually doing it"

You know, that would be a fun one to give CS students who think they know their stuff.

GreeBo

what are your two different ways?
I just did it there now...take a minute or two to get your head around the recursive logic, but it is much more elegant programming.

I got bored and decided to allow the class to take in any number of factors and return the result..still using recursion
ahh double recursion...:o

plenderj

I also toyed for a while in my head of doing a version whereby you divide by the reciprocal, but never bothered actually testing it...

PhantomBeaker

Um, there's iterative addition, which is the easy one. It's really simple.

Then there's this one (done in C):

int multiple(int n, int f){
  if( f < 0){ 
    f = -f;
    n = -n;
  }
  return (f == 0) ? 0 : (multiple(n, f >> 1) << 1) + ((f%2) ? n : 0);
}

I just remembered that one off the top of my head (i.e. not cut/paste).

Yes, it's demon code, but it's fun.

Aoife

Hobbes

plenderj wrote:

Now do it recursively

I thought recursion was bad?

plenderj

Who ever said recursion was bad?

GreeBo

PhantomBeaker wrote:

Um, there's iterative addition, which is the easy one.

you said you had 2 ways of each (so two methods of recursion and two methods of iterative)
What are your two methods?
I presume you have two ways of performing multiplication, then each implemented eithe recurisveuly or iteratively....

Hobbes

plenderj wrote:

Who ever said recursion was bad?

Well in Java for example it has a much larger memory overhead then using normal looping expressions. It is more prone to stack crashes if not coded correctly where as an expression loop isn't.

Thats not to say it shouldn't be used, but you should be aware of what the conditions are.

Its a bit like the StringTokenizer and the split() method.

PhantomBeaker

GreeBo wrote:

you said you had 2 ways of each (so two methods of recursion and two methods of iterative)
What are your two methods?
I presume you have two ways of performing multiplication, then each implemented eithe recurisveuly or iteratively....

Yes. the iterative addition one is just a for loop. I'm not writing that. It also takes log(f) (where f is the factor) time. (I'll leave that as an exercise to the reader )

The recursive version of that cuts the running time to O(log f).

int addmultiple(int n, int f){
        if(f < 0){
                f = -f;
                n = -n;
        }
        if(f == 0){ /* I'm pretty sure I can drop this, but hey.*/
                return 0;
        } else if (f ==1){
                return n;
        }
        int half_value = addmultiple(n, f/2 );
        return ((f % 2)? n : 0) + half_value + half_value;
}

The other method is essentially how hardware does it (although hardware has a couple of optimisations to that algorithm). As I posted the recursive version above, I'm going with the iterative version.

int b_multiply(int n, int f){
        if(f < 0){
                f = -f;
                n = -n;
        }
        int answer = 0;
        int counter =0;
        while(f > 0){
                if(f % 2){
                        answer += (n << counter);
                }
                counter++;
                // printf("f was: %d, f is now: %d\n", f, (f >> 1));
                f = f >> 1;
        }
        return answer;
}

Once you're not scared of the << and >> operators, and once you realise that shifting a bit up by one means that it's multiplying it by two, and shifting down by a bit is dividing by 2, it should all make sense.

In this case, the recursive and the iterative both have the same running times (O(log f)), so even though the recursive is prettier to look at, this actually takes up less memory, and doesn't use up as much of the stack. Plus it's easier to debug.

Smoggy

Show off's

Can you do it via memory shifts ?

edit :

I had no idea how to do this , but after 10 mins playing around , it turns out I do (vb.net)

Private Function Multiply(ByVal no1 As Integer, ByVal no2 As Integer) As Integer

Dim Value As Integer

If no1 > 0 Then

Value = Value + no2

no1 -= 1

Value = Value + Multiply(no1, no2)

End If

Return Value

End Function

edit edit : Just saw that Phanton did it with bit shifting, now thats definatly above me:D

Hobbes

In vb to bit shift I think you do...

value ^ 2

Been a while. that might be wrong.

plenderj

That'll just square a number

Smoggy

I'm not 100% sure myself , but I think 2003 uses << and >>

fatherdougalmag

Plenty of entertainment with the Gimpel 'Bug of the Month': http://www.gimpel.com/html/bugs.htm

yossarin

The technical test in my work is a debugging test.

we have a few java classes and you had to be able to step through the program and debug it.

its stuff like thinking to print and undertand a stack trace, and understanding classpaths, libraries and class herarchies. Its made to not be solvable in the time we give you (30 mins, took me 38).

One test i was given in an interview for inmarsat was a pointers C++ test. they'd show you a few lines of code using varibales and pointers and ask you what a particular variables value would be.

I've had a few interviews where i've been struggling though somthing and the engineer has solved it him/herself without meaning to - just while wanting to help.