Orthogonal Defect Classification

Orthogonal Defect Classification ( ODC ) is a model for the classification of errors in software development, which was developed at IBM. By the classification of errors according to certain categories and attributes, interrelations between the development process and the cause of errors can be highlighted. Based on the findings they suggest measures for the improvement of the development process can be derived.

  • 3.1 Root Cause Analysis
  • 3.2 Statistical Defect Modeling

Concept

ODC belongs to the class of root cause analysis (root cause analysis ) and is a technique from the field of software engineering. In contrast to classical root cause analysis in this case, the individual errors are not detailed and time-consuming analyzes, but semantically classified by means of predefined categories.

Error classification

In the fault classification with ODC, the error is classified twice - upon detection of the error according to the criteria of the so-called " Opener Section"; after clearing then again after which the "Closer Section". In the opener section the focus is on the circumstances that led to the error - so they can be reproduced later. In Section Closer the cause is determined, which have led to the error. The opener Section offers three categories, Closer Section five.

Section opener

The opener Section consists of the categories:

  • Activity: When was found an error? This is not the date meant, but the date in the development process and the product life cycle.
  • Trigger: As the fault was detected? Here, the trigger should not be confused with the symptom, ie the effect of an error or the misbehavior of a program.
  • Impact: What are the implications for the customer or could arise if the error had been discovered only at the customer in terms of ISO 9126?

Closer Section

The Closer Section consists of the categories:

  • Qualifier: Specifies whether the correction of an error due to: missing, wrong or liquid source was.
  • Age: Specifies whether the error in a new, base, rewritten ( refactored code) or fehlerbereinigtem code branch occurred.
  • Source: Specifies whether the error is within the source code of their own development, reuse ( reuse), porting or a third party vendor provides.

Possibilities of evaluations with ODC

The Orthogonal Defect Classification as a representative of software metrics is by its classification of fault data to be able to create on the basis of an ODC dataset evaluations regarding the particular software development process. Here, the ODC concept provides for a total of two methods by which the reliability and quality of a test and development process for software can be analyzed and evaluated:

Root Cause Analysis

As mentioned at the beginning, the RCA belongs to subgroup analysis of the cause of an error in software engineering. The ODC RCA distinguished in that it can perform not only much more efficient and faster, but also more accurate compared to other methods of root cause analysis. This is due to the structuring of the respective software fault case in the form of ODC attributes Opener and Closer Section. In the practice of an ODC RCA values ​​to the ODC attributes of the opener and Closer are analyzed and evaluated. By considering the ODC values ​​and interrelated ODC attributes of the RCA implementers can locate and determine a potential cause of the error in more detail. Despite the performance gain by the ODC - RCA, the general method of the RCA is classified as a relatively complicated and time-consuming. For this reason, a ( ODC ) RCA as possible by an experienced software developer and are primarily performed for higher priority or related software errors.

Statistical Defect Modeling

In contrast to the ODC RCA, in which only a single software error is considered in more detail, relates the SDM a component of a whole ODC data in the analysis. In SDM, the software error in its structure ODC - form machine analyzed by either the ODC attributes of the Opener and Closer Section one another or with other determinants, such as development period, product components, projects, etc., are placed in relation. The results of these evaluations do not like the ODC - RCA information on the origin of a single software error, but represent the actual state of the development process in terms of reliability and quality of the respective test or development phases generally graphical form

To illustrate the ODC data analysis as part of the SDM diagrams or even so-called Inferenzbäume be used. Due to the structured form of the ODC dataset these charts and Inferenzbäume can be automatically generated and evaluated.

Difficulties in the use of ODC

As it is with the ODC is a process carried out by people in the error classification, there is a risk that due to differences in subjective views the classification of an error leads to differences. This can be positively influenced by a specific adaptation of the attributes to the particular circumstances in the development of each category in the opener and Closer Section. Therefore, the individual attributes of a category should be orthogonal to each other so as to minimize the room for any interpretation by the individual.

Code refactoring Orthogonality
624606
de