This work has been supported by the European

This work has been supported by the European Social Fund within the project «Support for Doctoral Studies at University of Latvia» . Software Self-Testing Edgars Diebelis Prof. Dr. sc. comp. Jānis Bičevskis

Self-Testing Ability to execute stored test examples to ensure that software functionality is correct. Testing approach ◦ Manual testing; ◦ Automated testing with testing tools; ◦ Self-Testing. 2

Content Problem statement. Concept of Self-Testing. Implementation of Self-Testing. Comparison of the Self-Testing Concept with Conventional Testing Support Tools. Efficiency Measurements of Self-testing. Conclusions. 3

Problem Statement „Computing systems’ complexity appears to (Kephart, J. , O. , Chess, D. , M. ). IBM Autonomic Computing Manifest, 2001 be approaching the limits of human capability” ◦ outlined four key features that characterise autonomic computing. Smart Technology approach, 2007 ◦ identified seven types of the smart technology; ◦ self-testing is one of the type of the smart technology. 4

Concept of Self-Testing

Self-Testing Functionality Self-testing contains two components: ◦ Test cases of system’s critical functionality to check functions, which are substantial for system using; ◦ Built-in mechanism (software component) for automated software testing (regression testing) that provides automated running of test cases and comparing test results with standard values. 6

Test storage mode 1. 2. 3. Test storage file (XML) 7

Self-Testing mode 1. Test storage file (XML) 2. Self-Testing tool 6. 4. 5. Test storage file (XML) 3. 8

Test Points Test point is a programming language command in the software text. Test point ensures that: ◦ particular actions and field values are saved when storing tests; ◦ the software execution outcome is registered when tests are executed repeatedly. By using test points, it is possible to repeat the execution of system events. 9

Test Point Types Test point types: ◦ ◦ ◦ ◦ input field test point; comparable value test point; system message test point; SQL query result test point; application event test point; test execution criteria test point; self-testing test points; etc. 10

Implementation of Self-Testing

Build of Self-Testing tool modules: ◦ Self-Testing module; ◦ Self-testing test management module. 12

Self-Testing in use 13

14

Comparison of the Self-Testing Concept with Conventional Testing Support Tools

Testing Tools Selection Automated Testing Institute (ATI) opinion. “ATI Automation Honors ” awards. Since May 2009, the ATI has been publishing its magazine Automate Software Testing. Website has a list of 716 automated testing tools. ATI annual conference on automated testing (Verify/ATI). ~8000 registred users. 16

Testing Tools Test. Complete Fit. Nesse Ranorex T-Plan Robot Rational Functional Tester HP Unified Functional Testing Software Selenium the best commercial automated functional testing tool in 2010 the best open source code automated functional testing tool in the NET subcategory in 2010 the best commercial automated functional testing tool in the NET and Flash/Flex sub-categories in 2010 the best open source code automated functional testing tool in the Java subcategory in 2010 in 2009 and 2010, it was a finalist among the best commercial automated functional and performance testing tools the best commercial automated functional testing tool in 2009 the best open source code automated functional testing tool in 2009 and 2010 17

Criteria for Comparison Testing method (TM) Test automation approach (TAA) Test automation framework (TAF) Testing level Test recording and playback Desktop applications testing Web applications testing Services testing Database testing Testing in production environment System user can create tests Simultaneous running of several tests Performing simultaneous actions Identifying the tested object Test result analysis Test editing Screenshots Control points Object validation Object browser Test log Test schedule planner Identification of the end of command execution Plug-ins and extensions etc. 18

Comparison 19

Comparison II 20

Self-Testing further development New Test automation frameworks. Test editor and log. Object validation and object browser. Load, stress and other testing levels. Web applications and services testing. Additional platforms. Plug-ins and extensions. Integration with external environment testing. 21

Self-Testing advantages White box testing. Testing in production environment. Users without in-depth IT knowledge to define and run test cases. Testing external interfaces. Perform system testing without specific preparation for running the test. 22

Efficiency Measurements of Self -testing

Efficiency Measurements Retrospective analysis of incident notifications in a real project. It is not possible to apply and compare two different concepts in the same conditions. Analysis of incident notifications (1, 171 in total) in the CSAS in the period from July 2003 to 23 August 2011. Subjective opinion; however, the high number of incident notifications and the statistics do reflect trends. 24

Statistics of Incident Notifications Type of Incident Quantity % of total Hours % of total Duplicate 68 5. 81 23. 16 0. 47 User error 43 3. 67 67. 46 1. 37 Unidentifiable bug 178 15. 2 1011. 96 20. 52 Identifiable bug 736 62. 85 3293. 74 66. 79 Improvement 102 8. 71 241. 36 4. 89 Consultation 44 3. 76 293. 92 5. 96 1171 100 4931. 6 100 Total: Improvement Consultation 8. 71% 3. 76% Duplicate. User error Unidentifiable 5. 81% 3. 67% bug 15. 20% Identifiable bug 62. 85% 25

Bugs Unidentifiable by the Self-testing Bug type Quantity External interface bug % of total 5 2. 81 12 6. 74 7 3. 93 25 14. 04 Simultaneous actions by users 5 2. 81 Requirement interpretation bug 41 23. 03 Specific event 83 46. 63 178 100 Computer configuration bug Data type bug User interface bug Total: External interface bug 2. 81% Specific event 46. 63% Computer Data type bug configuration 3. 93% interface User bug 6. 74% 14. 04% Simultaneous actions by users Requirement 2. 81% interpretation bug 23. 03% 26

Bugs Identifiable by the Self-testing Quanti ty Test point % of total Hours % of total File result test point 59 8. 02 150. 03 4. 56 Input field test point 146 19. 84 827. 14 25. 11 Application event test point 105 14. 27 364. 24 11. 06 Comparable value test point 28 3. 8 93. 53 2. 84 System message test point 11 1. 49 58. 84 1. 79 SQL query result test point 387 52. 58 1799. 96 54. 65 736 100 3293. 74 100 Kopā: SQL query result test point, 52. 58 File result test point, 8. 02 System message test point, 1. 49 Entry field test point, 19. 84 Application event test point, Comparable 14. 27 value test point, 3. 8 27

Conclusions

Conclusion self-testing approach offers not only options equal to those offered by other globally recognized testing support tools; moreover, self-testing additionally offers options that other testing tools do not possess: ◦ ◦ testing external interfaces; testing in production environment; testing with the white-box method; possibility for users without IT knowledge to capture tests. Testing support is part of systems developed and it is available throughout the entire life cycle of software. 29

Conclusion II Self-testing changes the testing process by considerably broaden the role of the developer in software testing. Self-testing requires additional work to include the self-testing functionality in the software and to develop critical functionality tests. Self-testing saves time by repeated (regression) testing of the existing functionality. Implementation of the self-testing functionality is useful in incremental development models, in particular in systems that are gradually improved and maintained for many years. 30

Thank You! 31