Sunday, 19 January 2014

The Fallacy of Omniscience by Proximity



I was reading about President Hollande's apparently complex love life, and a journalist who used to live in France was asked his opinion on the issue. He responded that although he used to live there, those who wanted his opinion assumed that this meant he knew all about the issue and had a groundbreaking insight into the story. 

Turns out, he hadn't lived in France for years, so was about as wise as the rest of us. Those seeking his opinion had fallen for the wonderfully named:


'The Fallacy of Omniscience by Proximity'

It means (in a general sense) you are/used to be in close proximity to subject, so it stands to reason you must know all about it right?

After pondering this lovely phrase for a while, I realised it has a relevance for us testers. 

Firstly, often, testers are a respected oracle for a domain and/or an application. You know that person who 'has been here forever and knows everything about X.' That is literally not true, and reinforces that all oracles for testing (even our fellow testers) are fallible. I still see (and participate in) this behaviour often.

Secondly, when you have tested an application I have found that it is assumed you have knowledge of every major and minor path through the functionality. As a tester, consider this statement, perhaps by another product stakeholder:


'Well, you tested it! Why didn't you know that when I do Y it triggers behaviour Z?!?'

We know of the impossibility of complete testing (our stakeholders may not) but there can be an assumption that testing brings 'complete knowledge' of an application. This is a classic example of the fallacy in question, and has its roots in a fundamental misunderstanding about the purpose of testing that still permeates today.

For now, I'm delighted to find a new way to describe a problem I ponder often.

No comments:

Post a Comment