BomberMouse
Member
I agree, but a good interview analyze how you tackle the problem instead of the solution.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Programming |OT| C is better than C++! No, C++ is better than C
- Thread starter Kikarian
- Start date
I agree, but a good interview analyze how you tackle the problem instead of the solution.
Elfforkusu
Member
Regrettably, the real-world tested process of "research/play with problem for 15-30 minutes until you have a good feel for all the possibilities, then implement something" doesn't translate very well to the standard interview "write pseudocode of the first thing that pops into your head, because the next dude will be in here in 30 minutes".
It's a remarkably trite process, IMO. Luckily I'm employed, but every time I see a neat opportunity and jump back into the "game", I'm reminded how jarringly disparate interview questions and real world problems are. The process is the equivalent of judging the ability of endurance runners by having them sprint 100 meters.
I couldn't disagree more with your views on the interview process, in theory and in practice.
TheFeedingHand
Member
There's interviewers that ask people to code quicksort?
BomberMouse
Member
Yes, I got asked that on my last interview. Since it was 8+ years since I saw it I obviously blew all the little details.
Syracuse022
Member
Honestly, you sound like exactly the type of person who makes me glad that programming jobs lend themselves so well to a more quantitative interview process than most other fields.
I just spent about half an hour not realizing that:
means currentState will never change. Still grasping to about 1 millimeter of confidence with the fact I was spacing out after a surgery, but what the hell. I don't feel very smart right now, lemme tell you.
Code:
if (!canContinue()) {
currentState = PlayStates.GAME_OVER;
}
currentState = PlayStates.MOVING_PIECE;toddhunter
Member
Don't worry I've lost many an hour to something like:
Code:
if (a == b || temperature > 36);
doSomething;This is probably not the place for a question like this but oh well it seems computery enough.
I've been taking an architecture class and I must explain why the future of engineering research is not in CPUs but in Memory(in the I/O sense). Now I know that the reason why we've slowed optimization of CPUs, but I'm not exactly sure why we're shifting research to I/O access.
Don't really need a long explanation, but if someone could point me to a paper that could help me better understand what's going on.
Edit: took me a few hours, but I've found a couple of white papers to get me going. Forget I ever existed plox.
I've been taking an architecture class and I must explain why the future of engineering research is not in CPUs but in Memory(in the I/O sense). Now I know that the reason why we've slowed optimization of CPUs, but I'm not exactly sure why we're shifting research to I/O access.
Don't really need a long explanation, but if someone could point me to a paper that could help me better understand what's going on.
Edit: took me a few hours, but I've found a couple of white papers to get me going. Forget I ever existed plox.
I make so many dumb mistakes, it's just ridiculous. The worst thing about it is that I rarely even "discover" them, I just have to write the block from scratch and if it works, compare them line by line.
Freyjadour
Member
It is fucking bonkers that I cannot install VsVim in Visual Studio Express.
Fucking. Bonkers.
Fucking. Bonkers.
Dreamspark, fuck yeah.
Open-reference 2 hour email test covering algorithms, linear algebra, STL and C++ programming practices in C++ and Visual Studio. This is for an entry-level game programming position.
What can I expect and how do I go about preparing for this? I'm really rusty on those things atm. Sure, it's open reference, but I want to play it safe.
I have my textbooks besides me: Introduction to Programming with C++, Essential Mathematics for Games, Data Struct and Alg for C# and Discrete Mathematics. My Google-fu is up to par but I want to know if there's anything I'm missing.
What can I expect and how do I go about preparing for this? I'm really rusty on those things atm. Sure, it's open reference, but I want to play it safe.
I have my textbooks besides me: Introduction to Programming with C++, Essential Mathematics for Games, Data Struct and Alg for C# and Discrete Mathematics. My Google-fu is up to par but I want to know if there's anything I'm missing.
Freyjadour
Member
Vi and Emacs support should really be standard in Visual Studio anyway. I know I can bind gvim, but losing Intellisense, etc. makes things overly irritating.
Most of the time it's horseshit imo. I've gone through the whole process of memorizing implementations of data structures & algos for interviews. Which after landing the job, have never had to be written from scratch 99% of the time. If they did, you basically figure out how to implement it, usually from the book or the internet and code it to work.
Memorizing the stuff doesn't prove to me, or anyone really that you can program worth a shit imo. It proves that you can memorize things, which is important, but it's not the most important indicator of someone who can do the job imo.
I think it's much more important to know how to identify, work through, and eventually solve problems. What data structures and algos would you use to solve the problem efficiently? Use pseudocode to solve a problem, bonus if you can do it in the language(s) listed on your resume. Reveal what you are thinking as you work through the problem. Etc...
Anymore, I think implementing ds & algos should be a detail that is abstracted out in the interview, simply because it's a detail that is frequently abstracted out in our daily jobs.
I would rather listen to someone tell me why they would use a bubblesort over a quicksort to solve a particular problem, rather than someone who can jank out a non-production implementation. It's much more revealing that imo.
Granted we are speaking of likely entry-level positions, so coding skills are still going to have to be refined. Mid and senior level position would likely have to be more specific and in-depth.
Elfforkusu
Member
While I haven't worked in a non-technical field (apart from random jobs in college, which I don't count), I can only imagine it's even worse. What I'm opining about is the inadequacy of the quantitative test, not that there shouldn't be one or that it would be better to not have one. It's certainly better than nothing.
"trite" was a bit harsh, but I'm curious about your disagreement -- at least as far as the practical application goes (in theory all sorts of things that don't work, do).
Essentially what I'm saying is that the interview process is suboptimal (this cannot be anything but true: How can I judge which of ten candidates will be the best fit after 6 months without having them work for 6 months?), and at most companies it seems to be suboptimal in just the right way so as to drive me nuts. I should add that this is hardly the case at all places. I interviewed on-site at a fairly famous tech company a few months ago, and I came away pretty damn impressed by their interview process. Of course, I fumbled away my chance at a job offer because I'm an idiot, but... so it goes.
Back to the runners analogy: having all my candidates run the 100 meters does a few things. It eliminates the out of shape ones, the ones who can't run, and realistically, I can use how they ran the 100 to guess whether they might be able to adapt to distance running. It's not a worthless metric. But at the same time, I'm going to miss some of the guys who suck at sprinting and just happen to be great long distance runners. It's not the end of the world, but it's not great either. Those are the guys we want!
toddhunter
Member
Yeah it isn't perfect, after all, if anybody ever actually implemented a sorting algorithm in the real world they would probably be fired. But this is why it shouldn't be part of the whole process either. If anything whenever I have asked technical questions, I'm not really even paying attention to their answer (because "I would just google it and copy paste" is probably the correct one), it is more about how they go about answering it and if they seem to be able to "figure things out".
It is all part of the puzzle and yes there are certainly right and wrong ways to go about it.
krameriffic
Member
So I spent 6 months doing programming as a stopgap at the company I was hired at out of college while the contract I waited on was in limbo. I didn't think I'd like it, the program had a really bad reputation, and they threw me straight into the deep end of the pool. Now that the contract came through, I moved onto that program doing hardware/SA stuff... and it's boring and I hate it. For all the flaws of the previous program, it grabbed a hold of me and forced me to focus and work hard in a way that I don't think I ever had before, and I quite enjoyed it. Then I get thrust into a related field and I'm bored out of my goddamn mind.
I'm playing with the possibilities of how to make the move back into software. I've only been with the company for around 7 months now, so it feels too early to start actively looking for something else. But at the same time, the month I've spent doing hardware/SA has been almost a complete waste of time, and I hate the idea of having to spend another year or more doing it before I can make the move without looking like someone who is hopping around jobs too frequently. Plus, it is only making me more desirable and qualified for more of the same shit rather than contributing to a good programming skillset (there's only so much scripting an SA can do).
Anyone have some advice? Not sure if this is quite the right place for it, but it seems close enough.
I'm playing with the possibilities of how to make the move back into software. I've only been with the company for around 7 months now, so it feels too early to start actively looking for something else. But at the same time, the month I've spent doing hardware/SA has been almost a complete waste of time, and I hate the idea of having to spend another year or more doing it before I can make the move without looking like someone who is hopping around jobs too frequently. Plus, it is only making me more desirable and qualified for more of the same shit rather than contributing to a good programming skillset (there's only so much scripting an SA can do).
Anyone have some advice? Not sure if this is quite the right place for it, but it seems close enough.
Freyjadour
Member
I'm in a pretty similar situation career wise. Took the SA job because it was good money/benefits with a good company, but absolutely is not what I want to do for a living. I've been there a little under two years trying to break into their (smallish) internal dev group. Worked on a project or two but was pulled off as the SA group was downsized. I've pretty much been told it isn't in the cards for the time being (aka forget about it) due to my FTE status and experience in the SA role. Communication from management about all of this has been terrible and the last year has been an especially frustrating rollercoaster of opportunity and disappointment.
I get where you're coming from about boredom. I'm a very driven person, but I have no will to improve myself in the role that I'm in anymore. Very much out of character for me.
I'm at the point now where I'm readying myself to find something else. Not looking forward to the decrease in pay and benefits, but I am so ready to care about what I do for a living again.
The last 2 years feel like a giant void career-wise, and I think that will look worse in an interview than a small amount of job hoping. If you're anything like I am, I think that showing genuine enthusiasm about making a positive change, especially early in your career, should be met with understanding. I really wish I'd made a change at 7 months instead of holding out another year for a chance with this company.
What should your program do when it's run?
I've never encountered that before, but I would ensure my environment is setup so I could quickly write tests.
I've never encountered that before, but I would ensure my environment is setup so I could quickly write tests.
Randolph Freelander
Member
So far, you've written what barely amounts to part of a requirements statement. Care to expand more? What specifically is giving you problems?
Very start of it actually, I understand the constructors and such. But labeling what I need would be private and public. Basically the logic in initially breaking it down
JeTmAn81
Member
Randolph Freelander
Member
What I mean is... breaking what down? This is your entire problem description:
"Needs to be surface area and volume, with three private members of type double, and two constructors"
What needs to be surface area and volume? Are those your classes, or are those properties of a class? Is that surface, area, and volume, or surface area and volume? If the former, do those correspond to the members you reference, or something else? Your problem statement is incredibly vague. You didn't even mention the language, for instance. When seeking help in a public forum, understand that we're not mindreaders.
If you had asked your question on Stack Overflow, for example, you would probably already have 12 downvotes and your question would be closed as "not a real question." Bad times had by all.
Correct. To explain it in really rough terms, unless the if is followed by {} brackets, it only applies to the next statement ending in ;. In this case, that's an empty statement, so it applies to nothing.
That's one of several reasons why many style guides suggest always using {} brackets for if statements.
krameriffic
Member
I have a friend who works in the same industry, same area, similar programs, but for a smaller company (I'm with one of the really big juggernauts). Their pay and benefits are far better, but it's tougher to get in the door. He could be my way in, but it's a question of whether or not I've been with this company for long enough to pursue such a "team change". I feel like the one year mark would be a more appropriate time, and hell, it's only a few months away.
Any thoughts on Crucible? It looks like a pretty sweet code review tool. I'm really sick of slowly diffing files one by one from an archaic version control system.
It looks like git is in my future. I'd have chosen mercurial as it seems more human-friendly, but either way it's going to be an improvement. And an opportunity to start integrating better tools.
It looks like git is in my future. I'd have chosen mercurial as it seems more human-friendly, but either way it's going to be an improvement. And an opportunity to start integrating better tools.
String streams! I'm not sure why my code isn't working and is returning an error
Looked this up online and for some reason it refuses to compile correctly. Instead it's telling me this:
Code:
#include <iostream>
#include <string>
#include <fstream>
using namespace std;
int main()
{
int num = 1;
stringstream ss;
ss << num;
string result = ss.str();
cout << result << endl;
return 0;
}Looked this up online and for some reason it refuses to compile correctly. Instead it's telling me this:
Ooh ok then. For some reason all the examples I found online didn't include that.
We use mercurial and crucible at work. I love mercurial.
Crucible is OK. The interface is pretty cluttered. It can be hard to find what you're looking for. I'm sad to admit that we don't use it frequently enough for me to become familiar and the trivial UI annoyance to disappear. The functionality seems fine to me, but again I haven't used it heavily.
Definitely a million times better than manually diffing individual files.
IIRC the licensing is per user that shows up in your commit history. We only had like 5-6 real people, but that amounted to like 12 names in the history due to using slightly different names on older computers etc. Their licenses are 10 users for cheap or 100 users for megabux we didn't have. So we had to actually make new repos and fix the users in the history as well as manually clean out crucible's database. Took about a day. I could be remembering some/all of this incorrectly.
Lasthope106
Member
Trying to get back into C++ once again, but I'm having a slight problem with the following code. When I try to go through the loop a second time cin, reads the EOF character. Any way to avoid this?
edit: Nevermind, I had to add cin.ignore() after I clear the stream.
Code:
#include <iostream>
#include <stdexcept>
#include <string>
#include <vector>
using namespace std;
int main()
{
const int NUMBER_OF_HOMEWORKS = 5;
cout << "This program will keep track of a list of student and their grades." << endl;
cout << "If no more students to enter, type \'done\'." << endl;
vector<string> studentsNames;
vector<double> studentsFinalGrade;
while(true)
{
cout << "Enter the student name" << endl;
string name;
cin >> name;
if (name == "done")
break;
cout << "Enter " << name << " midterm and final grades:" << endl;
double midterm, final;
cin >> midterm >> final;
cout << "Enter the grades for all the 5 homeworks assigned this semester"
"/n separated by spaces. When finish, enter the EOF character (Ctrl-z)." << endl;
double count = 0, sum = 0;
double grade;
while (cin >> grade)
{
sum += grade;
count++;
}
if (count != NUMBER_OF_HOMEWORKS)
{
throw invalid_argument("The grades for the homework of this student were entered incorrectly");
}
double homeworkGrade = 0;
if (count != 0)
{
homeworkGrade = sum / count;
}
double finalGrade = 0.4 * final + 0.2 * midterm + 0.4 * homeworkGrade;
studentsFinalGrade.push_back(finalGrade);
cout << "The student " << name << " scored a " << finalGrade << " final grade in the course." << endl;
cout << endl;
cin.clear();
}
cout << endl;
if (studentsFinalGrade.size() == 0)
{
cout << "No students were entered" << endl;
}
else
{
for (int i = 0; i != studentsNames.size(); ++i)
{
cout << studentsNames[i] << " " << studentsFinalGrade[i] << endl;
}
}
return 0;
}edit: Nevermind, I had to add cin.ignore() after I clear the stream.
Mr_Appleby
Member
was getting drunk with a friend last night when he decided to set up his old Vic-20 that had been sitting in a box for 25 odd years. i hadn't touched a pre-amiga commodore computer in about that long myself but it all came back in seconds. spent the next hour hacking up a version of Snake in CBM basic. was crazy fun. makes me wanna dig up one of my families old C64's and get my code on.
smokeandmirrors
Banned
Tempted to take a graduate level "intro" computer science course as an elective next semester. Obviously even though it's a supposed "entry" level, it assumes you know math and logic as it's for a scientific computing graduate program (it's directed for math and other engineering majors so less hand holding). I talked to the professor teaching it and he said it's just straight programming in c++. My issue is it would put me at 18 credits so i'm looking at a full schedule, but I feel like having just a straight practical implementation of programming vs a more theoretical approach would be extremely beneficial and worth the extra work and lack of sleep.
Man, there are so many great final year / grad intro courses I would love to take.
One of my profs (pretty cool guy actually, assignments can be challenging but marking is quite good and you always wind up knowing a lot) mentioned that they're doing a retro games course. a little bit of history but also like... how did they manage to get pitfall to work on the atari?
a lot of optimization type stuff, it would be cool to go to.
One of my profs (pretty cool guy actually, assignments can be challenging but marking is quite good and you always wind up knowing a lot) mentioned that they're doing a retro games course. a little bit of history but also like... how did they manage to get pitfall to work on the atari?
a lot of optimization type stuff, it would be cool to go to.
If you're into that stuff, I'd recommend checking out Racing the Beam.
I wish I had more time. So many classes I would take purely for fun.
Thanks for your thoughts. They're really useful.
Is it usable without FishEye? Or is the FishEye integration integral to Cruciable?
Managed to answer 4 out of 5 questions on my programming test, I feel like a dumbdumb.
Also, I really spent way too much time reviewing matrix math when I should've just been looking at code.
This one in particular completely stumped me:
Does anyone have any insight to share on this?
Also, I really spent way too much time reviewing matrix math when I should've just been looking at code.
This one in particular completely stumped me:
Code:
//////////////////////////////////////////////////////////////////////////////////////////
//
// QUESTION 2: Write a template function below which returns the size (in bytes) of
// its template type without using the sizeof() operator. Use comments to explain your
// reasoning.
//
//////////////////////////////////////////////////////////////////////////////////////////Does anyone have any insight to share on this?
Just a thought, but maybe just shifting 0x80<<8 and count until it is 0?
ColtraineGF
Member
Well, originally I had no idea how to do so, but I did find one implementation on SO - it involves incrementing the address of the type pointed to by 1, and subtracting the original address to find out the size, like so:
(char*)(&obj + 1) - (char*)&obj
Doing &obj + 1 would give you the next offset of &obj in memory, moved forward by the size of 1 unit of obj, and subtracting from that value (converting the single unit of obj to however many bytes/chars) the original address gives you the size in bytes of obj.
How very neat, and it makes sense too. I'll have to remember that one.
(char*)(&obj + 1) - (char*)&obj
Doing &obj + 1 would give you the next offset of &obj in memory, moved forward by the size of 1 unit of obj, and subtracting from that value (converting the single unit of obj to however many bytes/chars) the original address gives you the size in bytes of obj.
How very neat, and it makes sense too. I'll have to remember that one.
Quiet a bit more elegant and constant than my solution, and even should even work with floating point types^^.
Fantastical
Death Prophet
This came in the mail today.
Yay, excited!
Yay, excited!
I've never used it without FishEye. Not really sure how it'd work. They seem kind of like one product to me.
Is it standard practice to use the library toString instead of creating your own toString? We were taught to use our own toString() method.