Question

Suppose that 100 errors were intentionally placed in a large software system before the system was subjected to final testing. Moreover, suppose that 200 errors were discovered and corrected during this final testing, of which 50 errors were from the group intentionally placed in the system. If the remaining 50 known errors are then corrected, how many unknown errors would you estimate are still in the system? Explain why.

Short answer

150 unknown errors may still be in the system, estimated using the ratio of initially found unknown errors to placed known errors.

Step-by-step solution

Identify Placed, Found, and Remaining Errors

The problem states that 100 errors were intentionally added to the system before testing. During testing, 200 errors were found, of which 50 were from the placed errors. Therefore, 100 placed errors - 50 found placed errors = 50 placed errors found remaining. This leaves us with 50 placed errors in the system, before correcting the remaining errors.

Calculate Total Errors After Remaining Known Errors

If the remaining 50 known errors (from the 200 discovered errors) are corrected, we need to identify the unknown errors. Since testing involved 200 errors and 50 were from the placed group, the remaining 150 (200 - 50) were originally unknown.

Estimate Remaining Unknown Errors

The initial 100 placed errors account for a subset of the total errors discovered, with 50 errors left after correction. Thus, any excess beyond the placed errors were initially unknown. Since 150 original unknown errors were found, this suggests the initial magnitude of unknown errors (unknown out of unknown errors found): 150 unknown / 50 placed known found = 3 times as many unknown as known errors. Therefore we estimate 150 unknown errors remain.

Key concepts

Software Testing

Software Testing is an essential part of the software development process. It involves evaluating a software application to ensure it meets the desired requirements and functions correctly without defects. There are various stages in the testing lifecycle including unit testing, integration testing, system testing, and user acceptance testing. Each stage plays a vital role in identifying and resolving defects that could negatively impact the user experience.
Testing methodologies can be broadly categorized into manual and automated testing. Manual testing involves human intervention, where testers manually execute test cases without the use of any scripting or automation tools. On the other hand, automated testing uses specialized software to control the execution of tests and compares actual outcomes with predicted outcomes. Both methods are crucial, and often a combination of both is used to create a robust testing strategy.

• Manual Testing - Effective for exploratory, usability, and ad-hoc testing scenarios.
• Automated Testing - Suitable for repetitive, regression, and load testing scenarios.

By detecting errors early in the development cycle, software testing helps to reduce the cost and effort required to fix issues at later stages. It also ensures the final product is reliable, performant, and ready for deployment.

Error Correction

Error Correction in software refers to the process of identifying, fixing, and verifying defects or bugs in a software program. This is an important step in the software development lifecycle, following the detection of errors during testing. In our given scenario, during final testing, 200 errors were found, and an important subset of these, intentionally planted errors from a group of 100, needed to be addressed.
Effective error correction involves several steps:

• Identification: Discovering errors through systematic testing or user feedback.
• Diagnosis: Analyzing the problem to understand its root cause.
• Fixing: Implementing changes to the code to resolve the defect.
• Verification: Re-testing the software to ensure the correction was successful and no new issues were introduced.

This process not only corrects the current defects but can also prevent similar errors in the future. Continuous error correction helps improve the quality and security of the software system.

Unknown Errors Estimation

Unknown Errors Estimation is a technique used to predict the number of potential defects remaining in a software system after a round of testing and error correction. In our exercise, we initially planted 100 errors, and during testing, 200 errors in total were found. Of these, 50 were known planted errors, leaving 150 as originally unknown.
To estimate remaining unknown errors, it is useful to compare the ratio of known errors found during testing to the unknown errors discovered. If 150 previously unknown errors were found and only 50 known planted errors were discovered, it suggests that the rate of finding unknown errors was higher than anticipated. This can lead us to estimate that approximately 150 unknown errors remain unaddressed, based on the ratio of known found to unknown discovered.
This estimation has vital implications:

• Project Planning: Provides insight into potential additional testing required before release.
• Risk Management: Helps assess the level of risk in deploying the software in its current state.
• Quality Assurance: Ensures that due diligence is applied in attempting to reduce defect levels.

While no estimation method is perfect, these calculations serve as a critical guide for further testing and development efforts.