Content deleted Content added
Manuel.oriol (talk | contribs) |
Citation bot (talk | contribs) Alter: title, template type. Add: chapter-url, chapter. Removed or converted URL. Removed parameters. Some additions/deletions were parameter name changes. | Use this bot. Report bugs. | Suggested by Headbomb | Linked from Wikipedia:WikiProject_Academic_Journals/Journals_cited_by_Wikipedia/Sandbox | #UCB_webform_linked 686/1156 |
||
| (82 intermediate revisions by 41 users not shown) | |||
Line 1:
{{
'''Random testing''' is a black-box [[software testing]] technique where programs are tested by [[random number generation|generating]] random, independent inputs. Results of the output are compared against software specifications to verify that the test output is pass or fail.<ref name="Hamlet94"/> In case of absence of specifications the exceptions of the language are used which means if an exception arises during test execution then it means there is a fault in the program, it is also used as a way to avoid biased testing.
== Overview ==▼
Random testing for hardware was first examined by [[Melvin Breuer]] in 1971 and initial effort to evaluate its effectiveness was done by Pratima and [[Vishwani Agrawal]] in 1975.<ref>{{cite journal|title=Probabilistic Analysis of Random Test Generation Method for Irredundant Combinational Logic Networks|first1=P.|last1=Agrawal|first2=V. D.|last2=Agrawal|date=1 July 1975|journal=IEEE Transactions on Computers|volume=C-24|issue=7|pages=691–695|doi=10.1109/T-C.1975.224289}}</ref>
In software, Duran and Ntafos had examined random testing in 1984.<ref>{{cite journal|title=An Evaluation of Random Testing|first1=J. W.|last1=Duran|first2=S. C.|last2=Ntafos|date=1 July 1984|journal=IEEE Transactions on Software Engineering|volume=SE-10|issue=4|pages=438–444|doi=10.1109/TSE.1984.5010257}}</ref>
The use of hypothesis testing as a theoretical basis for random testing was described by Howden in ''Functional Testing and Analysis''. The book also contained the development of a simple formula for estimating the number of tests ''n'' that are needed to have confidence at least 1-1/''n'' in a failure rate of no larger than 1/n. The formula is the lower bound ''n''log''n'', which indicates the large number of failure-free tests needed to have even modest confidence in a modest failure rate bound.<ref name=":0">{{Cite book|last=Howden|first=William|title=Functional Program Testing and Analysis|publisher=McGraw Hill|year=1987|isbn=0-07-030550-1|___location=New York|pages=51–53}}</ref>
Consider the following C++ function:
<syntaxhighlight lang="cpp">
int myAbs(int x) {
if (x > 0) {
return x;
}
Line 16 ⟶ 23:
</syntaxhighlight>
Now the random tests for this function could be {123, 36, -35, 48, 0}. Only the value '-35' triggers the bug. If there is no reference implementation to check the result, the bug still could
<syntaxhighlight lang="cpp">
Line 23 ⟶ 30:
int x = getRandomInput();
int result = myAbs(x);
assert(result >= 0);
}
}
</syntaxhighlight>
The reference implementation is sometimes available, e.g. when implementing a simple algorithm in a much more complex way for better performance. For example, to test an implementation of the [[Schönhage–Strassen algorithm]]
<syntaxhighlight lang="cpp">
int getRandomInput() {
}
Line 39 ⟶ 46:
long y = getRandomInput();
long result = fastMultiplication(x, y);
assert(x * y == result);
}
}
</syntaxhighlight>
While this example is limited to simple types (for which a simple random generator can be used), tools targeting object-oriented languages typically explore the program to test and find generators (constructors or methods returning objects of that type) and call them using random inputs (either themselves generated the same way or generated using a pseudo-random generator if possible). Such approaches then maintain a pool of randomly generated objects and use a probability for either reusing a generated object or creating a new one.<ref name="AutoTest"/>
== On randomness ==▼
According to the seminal paper on random testing by D. Hamlet
<blockquote>[..] the technical, mathematical meaning of "random testing" refers to an explicit lack of "system" in the choice of test data, so that there is no correlation among different tests.<ref name=Hamlet94>{{cite book|title=Encyclopedia of Software Engineering|year=1994|publisher=John Wiley and Sons|isbn=978-0471540021
==Strengths and weaknesses==
▲== Types of random testing ==
{{Unreferenced section|date=August 2014}}
Random testing is praised for the following strengths:
*It is cheap to use: it does not need to be smart about the program under test.
*It does not have any bias: unlike manual testing, it does not overlook bugs because there is misplaced trust in some code.
*It is quick to find bug candidates: it typically takes a couple of minutes to perform a testing session.
*If software is properly specified: it finds real bugs.
The following weaknesses have been described :
=== With respect to the input ===▼
*It only finds basic bugs (e.g. [[null pointer]] dereferencing).
* Random input sequence generation (i.e. a sequence of method calls)▼
*It is only as precise as the specification and specifications are typically imprecise.
* Random sequence of data inputs (sometimes called stochastic testing) - f.ex. a random sequence of method calls▼
*It compares poorly with other techniques to find bugs (e.g. [[static program analysis]]).
* Random data selection from existing database▼
*If different inputs are randomly selected on each test run, this can create problems for [[continuous integration]] because the same tests will pass or fail randomly.<ref name="so">{{cite web|url=https://stackoverflow.com/q/636353 |title=Is it a bad practice to randomly-generate test data?|website=stackoverflow.com|accessdate=15 November 2017}}</ref>
*Some argue that it would be better to thoughtfully cover all relevant cases with manually constructed tests in a white-box fashion, than to rely on randomness.<ref name="so" />
*It may require a very large number of tests for modest levels of confidence in modest failure rates. For example, it will require 459 failure-free tests to have at least 99% confidence that the probability of failure is less than 1/100.<ref name=":0" />
==
* undirected random test generation - with no heuristics to guide its search▼
* directed random test generation - f.ex. "feedback-directed random test generation"<ref name="PachecoLET2007">{{cite journal|last=Pacheco|first=Carlos|author2=Shuvendu K. Lahiri |author3=Michael D. Ernst |author4=Thomas Ball |title=Feedback-directed random test generation|journal=ICSE '07: Proceedings of the 29th International Conference on Software Engineering|date=May 2007|pages=75–84|url=http://people.csail.mit.edu/cpacheco/publications/feedback-random.pdf|publisher=IEEE Computer Society|issn=0270-5257}}</ref>▼
▲*
===Guided vs. unguided===
▲*
== Implementations ==
Some tools implementing random testing:▼
* [[QuickCheck]] - a famous test tool, originally developed for [[Haskell (programming language)|Haskell]] but ported to many other languages, that generates random sequences of API calls based on a model and verifies system properties that should hold true after each run. Check this [http://www.quviq.com/documents/QuviqFlyer.pdf QuviQ QuickCheck flyer] for a quick overview.▼
* [https://code.google.com/p/randoop/ Randoop] - generates sequences of methods and constructor invocations for the classes under test and creates [[JUnit]] tests from these▼
* [https://github.com/Datomic/simulant/wiki/Overview Simulant] - a [[Clojure]] tool that runs simulations of various agents (f.ex. users with different behavioral profiles) based on a statistical model of their behavior, recording all the actions and results into a database for later exploration and verification▼
▲Some tools implementing random testing:
== Critique ==▼
▲* [[QuickCheck]] - a famous test tool, originally developed for [[Haskell (programming language)|Haskell]] but ported to many other languages, that generates random sequences of API calls based on a model and verifies system properties that should hold true after each run
▲*
▲*
* AutoTest - a tool integrated to EiffelStudio testing automatically Eiffel code with contracts based on the eponymous research prototype.<ref name="AutoTest"/>·
* York Extensible Testing Infrastructure (YETI) - a language agnostic tool which targets various programming languages (Java, JML, CoFoJa, .NET, C, Kermeta).
* GramTest - a grammar based random testing tool written in Java, it uses BNF notation to specify input grammars.
<blockquote>Random testing has only a specialized niche in practice, mostly because an effective oracle is seldom available, but also because of
For programming languages and platforms which have contracts (
==
*
*
*
*[[Corner case]]
*[[Edge case]]
*[[Concolic testing]]
==
{{Reflist}}<!--<ref name=":0" />-->
==
*
*
*
{{software testing}}
[[Category:Software testing]]
| |||