4 ms·
The number one predictor of software bugs is organizational complexity: study
- tosh 7y ago> The distance to decision makers and the number of developers working on a project is clearly and unambiguously the issue that is the best predictor of future problems with a code base.
- ngneer 7y agoThe original tech report is from 2008. Anything newer?
- henryfjordan 7y agoThe study that's linked in the article replicating the results is from 2015
- cilefen 7y agoHere is Google’s take on bug prediction from a while back: https://google-engtools.blogspot.com/2011/12/bug-prediction-at-google.html?m=1 https://google-engtools.blogspot.com/2011/12/bug-prediction-...
- valerij 7y ago> Another shocking discovery for me personally, is that the only one that I've actually used myself - code coverage - has the lowest recall. In other words, almost all the issues it predicted turned out to not be real issues. sounds like survivorship bias to me
- drewcoo 7y agoI'd be willing to bet that it was sloppy coverage and those false positives were along component/team boundaries.
- drewcoo 7y agoTake Conway's Law into account and it's maybe not so surprising. https://en.wikipedia.org/wiki/Conway's_law https://en.wikipedia.org/wiki/Conway's_law
- yodon 7y agoThe original paper on which this blog post is based is actually an explicit experimental test of Conway's law (and it's also a paper which happens to have found significant validity for Conway's law in the data they analyzed).
- franciscop 7y agoCould a mod please update the title to make it less click-bait? "According to Microsoft, the number one predictor of software bugs is" sounds like all those side/bottom extra articles.
- dang 7y agoI took a crack at it, but the article has been flagged at this point so it's a bit moot.
- rambojazz 7y agoclickbait level: buzzfeed