Predicting the bugs

Recently I was going through the lecture video by Jim Whitehead wherin the author discusses about predicting the bugs using machine learning methods. The prediction is based on identifying those files and part of files which are more error prone and any change in that code is all set to create a bug. The author does not consider the developer's factor while predicting the bugs, but I personally feel that developer's factor should be considered since a developer creating a bug is prone to create a similar bug later even if its some other part of code. But the author's argument is that same developer can be writing a good code at one time and bad code at other. Predicting bugs is not a new thing in software industry, currently its done manually and the author wants to make it automated. Testing team will be keeping track of those parts of code which is more prone to errors and any modification in this code will need them to test it thoroughly. The teams also consider the developer's factor since many times its observed that code developed by particular devel is always more buggy compared to his counterparts.


  1. Jyothi12:53 AM

    Development and testing differs from person to person. Its difficult to categorize with 100% confidence. Developer can avoid same kind of mistakes if he/she remembers past history. I would say tester should be smart enough to predict the scenarios that might fail.

  2. Jyothi12:54 AM

    Forgot to say...Article is too good. Keep posting!!!!!