Tell me more ×
Programmers Stack Exchange is a question and answer site for professional programmers interested in conceptual questions about software development. It's 100% free, no registration required.
up vote -6 down vote favorite

We are a group of students who are working on a theory in software engineering (talking about the theory takes a lot of time so I just skip that).

Implementing the theory is impossible, due to technical limitations, but it can be proven on a paper logically.

We've been pushed to do a testing on it, so it can be proved that way too (although we bleieve that's not possible!), now:

  1. Basically, is it possible to test something like this?

  2. If it is, what type of testing should we use?

  3. I heard,its possible to handout a brief about it to some experts and asking about their opinion (not sure if that's true), is that a testing method? if it is, what does it called? and how exactly can be done?

share|improve this question
11  
Uhm ... I think you need to be a bit more specific about what your theory is. – Joachim Sauer Jul 9 '12 at 11:18
the testing phase can be considered a part of the theorizing process en.wikipedia.org/wiki/Scientific_theory because this can lead to observation – user827992 Jul 9 '12 at 11:31
to make it very simple, we are trying to prove that if we can't make sure that A does exist, its ok, as long as we know its been created from b+c+d, and we can prove that these 3 have been added to each other, an A should exist somewhere out there. – user363295 Jul 9 '12 at 11:36
Though I like this question very much, I'd suggest to close it and migrate it to CS.SE – haylem Jul 9 '12 at 11:51
2  
Here's a meta question: can you get an answer on SE to a question you won't ask? How the heck can anybody tell you whether you can test something if you don't tell them what that something is? How can they recommend a type of testing for an unknown hypothesis? If you can't be bothered to summarize your hypothesis, why should anyone help you? – philosodad Jul 9 '12 at 12:41
show 7 more comments

closed as not a real question by maple_shaft♦ Dec 26 '12 at 19:28

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, see the FAQ.

2 Answers

up vote 7 down vote accepted

You wrote

Implementing the theory is impossible

and

is it possible to test something like this?

so what exactly would you test if you don't have anything to test? Testing means that you have an implementation that you expose to various inputs and expect some particular output / behavior.

Proving is a different thing altogether. Keep in mind Donald Knuth's words: Beware of bugs in the above code; I have only proved it correct, not tried it.

share|improve this answer
I loved your answer :D, you are right i should have wrote it differently, ok its in-feasible to implement it, at the moment,and that was exactly the purpose of this research, we suppose to prove its correct, without any guaranty that it should work, – user363295 Jul 9 '12 at 11:51
Ok, so you mean proving, not testing. Please edit the question accordingly so you might get some useful answers. – Joonas Pulakka Jul 9 '12 at 11:56
up vote 3 down vote

If your theory makes falsifiable predictions, then it can be tested, to an acceptable level of statistical significance which will suggest your theory is indeed correct. For example, my theory is that tossing a coin is a random act. The prediction is that out of 100 coin tosses, the number of heads to tails will be approximately 50/50. I can test this prediction, and if the results do not deviate more than e.g., 5% then this is evidence that my theory is correct. Once you have gathered enough evidence, you can claimed to have informally proved your theory.

But perhaps you are after a mathematical proof? Again this is possible in code, assuming you have a theory which lends itself to this type of proof, but this is a rather specialised area.

share|improve this answer
Let me see if i got right, you mean if it can produce some results that can be tested, so i can arrange an experiment, and demonstrate the results statistically? well, as i mentioned above here is the idea: to make it very simple, we are trying to prove that if we can't make sure that A does exist, its ok, as long as we know its been created from b+c+d, and we can prove that these 3 have been added to each other, an A should exist somewhere out there, we don't know where is the A, and how can be measured, but we know that there is an A, because there is b,c and d. – user363295 Jul 9 '12 at 11:59
@user:your example is clear as mud. Are you saying that given bcd ==> A OR your hypothesis is given bcd ==> A but you aren't sure. In either case, it sure seems like you need more information to draw any conclusions. – Dunk Jul 9 '12 at 20:20
ok i've explained it above: alright, here is what we are trying to do, we are trying to categorized the functional requirements, under three main categorize, the inputs, outputs and the processes, that way we can create sth like a change between the inputs and the outputs, so we can trace them back to their origins, based on the inputs, it can be considered as a method to trace, but automating the process is not feasible at the moment, and we've been only asked to prove that such a thing can be true. – user363295 Jul 10 '12 at 8:44

Not the answer you're looking for? Browse other questions tagged or ask your own question.