Michael Jones, PhD

Full reference is: Chillarege, R., Bhandari, I., Chaar, J., Halliday, M., Moebus, D., Ray, B., and Wong, M. "Orthogonal Defect Classification -- A Concept for In-Process Measurements," IEEE TSE, vol. 18, no. 11 (Nov. 1992), pp. 943-956.

We like the classification scheme. I am particularly interested in which kinds of errors we think we will find with our tool. Then we can say more precisely how our tool will effect reliability using the seeded errors model. The error types are:

• function: something didn't get implemented. Requires a formal design change.
  
• assignment: a few lines of code, like initialization, didn't work.
  
• interface: errors interacting with other components. Device drivers etc.
  
• checking: didn't check data input values.
  
• timing/serialization: need improved management of shared and real time resources.
  
• build/package/merge: mistakes in libraries, version control etc.
  
• algorithm: effeciecy or correctness problems that can be fixed by changing the algorithm and nothing else.
  

They claim that timing/serialization errors will appear late in the design cycle. That's good because that's where we are aimed and errors found late are more expensive to fix. We'll spit into timing errors, serialization errors, algorithm errors and assignment errors. We don't do much with the rest.