Nov 2015
4 / 4
Nov 2015
Nov 2015
ilonek_1806
Nov '15
1

I think that validation of solutions does not work in some cases. There is no possibility to check what exacly happened in your program and with what input data it does not work. Maybe such guessing is supposed to be a fun..
Although, there are few examples of websites which provides also many different algorithmic problems from many categories, much more intuitive interface, embeded compilators which can be used before submitting the solution and what is the most important the possibility to check the input data and results of every validation test case. Would it be possible one day to upgrade SPOJ with such functionalities?

  • created

    Nov '15

  • last reply

    Nov '15
  • 3

    replies

  • 993

    views

  • 2

    users

  • 1

    link

leppyr64
Nov '15

Discarding my moderator hat here for a real post:
I disagree with the fact that this is "guessing". This site actually more realistically matches what happens in the "real world". When I am given a task at work, I need to ask questions and determine the requirements before I start to write code. After I'm finished writing my code I need to test it on a wide variety of situations to verify that it works. If I release it to the public and it doesn't work on some input that the user puts into it, it puts the whole business in jeopardy.

Imagine now that clicking "Submit" is equivalent to releasing your code to the public. Does it make more sense now?

This website is not a "coding" website. It is a problem solving website. That is why there are so many languages available. Code is just a tool to implement the solution to a problem.

Moderator Hat Back On:
If you are getting a wrong answer on this website it is more likely because you have handled the constraints of the input incorrectly and not considered possible test cases than the fact that the judge is not correct. If a judge is not correct, there will be many many comments mentioning that and people will not be getting accepted on it. If people are getting accepted on it, then the issue is with your solution. I suggest posting for help on the forum in the Problemset section, someone will likely help you.

ilonek_1806
Nov '15

In the real world we have debbuger's, which help us to find what exacly does not work in the program. What is more, we can receive from the client much more requirements and test cases than 1-4 as it is usually given in the problem decription.

leppyr64
Nov '15

Often in my experience working as an internal developer is that the requirements are often even less descriptive than the ones we get here. Then it is left up to me as the developer to determine actual requirements and limitations. At least here the constraints are given. If I release and fail the responsibility is mine.

Debuggers help track a problem after there is one, but it only helps after I have already failed.

The sample input is not expected to be representative of all possible inputs. It's just sample input. Part of solving a problem is determining all possible inputs and inputs that would stretch your solution.