Truth in Science and Quality

Positivism

Positivism in science is all about the objective and objectivity. This means there’s are separate, independent truths outside of anyone’s thoughts or beliefs. This truth can be observed, measured and analysed directly and accurately. All facts can be logically broken down and analysed without any feelings, bias or personal opinion. Just sticking to the facts. It’s called positivism because all theories about the truth of the universe can be “positively verified” with evidence. This leaves no room for personal interpretation, so everyone doing the same experiment will always see the same evidence in the same way and therefore form the same conclusion of truth. Anything abstract, invisible, cannot be measured or purely theoretical is rejected. (Robson, 2011a).

 

 

 

Post-Positivism

Post-positivism builds upon positivism by addressing some of its perceived issues. There is still an objective, independent truth of the universe however this truth cannot be objectively measured nor can evidence ever be observed directly. Objectivity should still be strived for as much as possible, but subjectivity will always play a part. Subjective and subjectivity refers to the person’s personal feelings, biases and opinions based on culture, life experiences and other things that only apply to people. This subjectivity will always distort the truth in some way and leave it open to some interpretation. The more research done, the closer one can get to perfect truth, but it will never be known completely or be 100% verifiable (Robson, 2011a) (hence the phrase “a theory can never be proven, only disproven”).

 

 

Constructionism and Interpretivism

Constructionist means there’s no independent, universal truth at all. Instead, everything is subjective. This means truth only exists in the minds of people and it is constructed based on their own personal feelings, beliefs and life experiences. People’s thoughts and opinions are what define truth, not anything observable in the real world. Therefore there is no directly observable evidence that can be measured and positively verified. Instead there are as many different truths as there are people to experience them. In this way, it’s opposite to the “realist” nature of positivism and post-positivism. Of course, the notion of separate truths is an independent, universal truth in itself so the philosophy is ironically self-disproving.

Relativism (also known as interpretivism) is similar except each person instead forms their own relative truth based on the real world and their experiences together. It advocates neither the objective nor subjective on its own but both which make up the truth (Robson, 2011b).

 

 

Coastline Paradox

How long is the coastline of Great Britain? Could you guess? Could you measure it? Perhaps look it up? The answer is: there is no answer. At least, no independent observable truth. Even though this is something in the natural world that can be measured, and people do measure it (see below), the answer is subjective and down to interpretation of what makes up a coastline. Coastlines are fractal, which means the smaller the units of length and more detailed the measurement, the longer the coastline gets (Mandelbrot, 1967).

This is important to know when testing, because a similar problem will be encountered every time a software product, system or service is tested for quality.

 

Abstract Relativity

Ideas of quality could certainly have been influenced by other people, products and the world around us making it subjective. Deming and Juran advocated quality as customer satisfaction, a purely subjective viewpoint. Standards and specifications, requirements and designs are also examples of someone’s subjective ideas. So quality isn’t objective. The software itself doesn’t have quality. Software can’t be pointed at and say this has good quality or this has poor quality. This is a fallacy of taking an abstract concept and making it physical and concrete:

Quality is not a thing and it is not built. To think of it as a thing is to commit the “reification fallacy”. Instead, quality is a relationship. Instead of “building” quality, it’s more coherent to say we arrange for it.

(Bach, 2009)

Quality isn’t absolute either. It doesn’t exist purely on it’s own. An assessment of quality cannot be performed independently, only in comparison to something or someone else. After all “quality does not exist in a nonhuman vacuum. Every statement about quality is a statement about some person(s)” (Weinberg, 1992). Windows 95 was a breakthrough operating system at the time, however compared to modern standards, it’s quality would be considered poor (relative). Not by everyone however, as some enthusiasts still get value out of old operating systems (subjective) (Basinger, 2006).  As a relationship, quality only exists relative to someone’s thoughts and ideas. This is known as the “relative” rule:

For any abstract X, X is X to some person

(Bolton, 2010a)

Thoughts, ideas, feelings and perceptions can change over time. Abstract concepts are therefore temporary. This also opens the time dimension to the relative rule. As with the definition of quality, all abstract concepts are relative to some person at some time (Bolton, 2010b).

While the truth of quality maybe subjective and relative, for the practical purposes of making software, an objective analysis is also required. There is still physical software which must be analysed to ascertain certain objective facts. The relative quality relationship comes from comparing the objective software to the subjective thoughts and ideas through a perception filter. Both the objective and the subjective together create the relative quality.

 

            ,:/+/-                      
            /M/              .,-=;//;-  
       .:/= ;MH/,    ,=/+%$XH@MM#@:     
      -$##@+$###@H@MMM#######H:.    -/H#
 .,H@H@ X######@ -H#####@+-     -+H###@X
  .,@##H;      +XM##M/,     =%@###@X;-  
X%-  :M##########$.    .:%M###@%:       
M##H,   +H@@@$/-.  ,;$M###@%,          -
M####M=,,---,.-%%H####M$:          ,+@##
@##################@/.         :%H##@$- 
M###############H,         ;HM##M$=     
#################.    .=$M##M$=         
################H..;XM##M$=          .:+
M###################@%=           =+@MH%
@################M/.          =+H#X%=   
=+M##############M,       -/X#X+;.      
  .;XM##########H=    ,/X#H+:,          
     .=+HM######M+/+HM@+=.              
         ,:/%XM####H/.                  
              ,.:=-.                     

Original ASCII art by Valve Corporation from Portal

 

• “I don’t like the cake” is a purely subjective viewpoint. Someone else may like it.
• “The cake is sweet” is a relative premise. The cake is bland when compared to a bag of sugar.
• “I don’t like the cake because it is too sweet” is a subjective-relative viewpoint. The feeling exists in relation to cakes from previous experience.
• “The cake has 200g of sugar” is an objective-absolute fact. Something that can be directly measured.
• “The cake has 100g of sugar more than the average” is an objective-relative fact. Something directly measured but only in comparison to other cakes.
• “I don’t like the cake because it is too sweet. It has 100g of sugar more than average compared to similar cakes” is a subjective-relative viewpoint with an objective-relative analysis.

For testers researching quality, this means considering all three of:

• Analysis of the physical software as it exists
• The ideas people have about the software and their immediate emotional response
• Comparison between the two and an assessment of how they both relate together

As the relationship between person and product grows in a positive way, so does its quality:

Summary Quality is a relationship between people, not an attribute of the software itself. Testers don’t communicate quality based on observed facts, rather the subjective viewpoint is required. Likewise, tester’s don’t just talk about quality in subjective terms either. There has to be empirical evidence or objective observed facts too. Quality isn’t absolute either, it only exists in comparison to something and someone else and their mental models. This means testing is both subjective, objective and relative requiring testers to focus on the whole relationship of software. As software quality is about people, relationships and their social and cognitive problems, software testers share a lot with real-world researchers and anthropology. This makes quality a social science.

Next:

• Quantitative vs Qualitative - Testing is journaling, note-taking and immersion resulting in a written evaluation of quality carefully supported by metrics where appropriate

 

Citations

• Bach, J., 2009. Quality is Dead #2: The Quality Creation Myth. [online] Satisfice. Available at: Link
• Bolton, M., 2010a. Done, The Relative Rule, and The Unsettling Rule. [online] DevelopSense. Available at: Link
• Bolton, M., 2010b. Quality is value to some person at some time. [online] Markus Gärtner. Available at: Link
• Mandelbrot, B., 1967. How Long Is the Coast of Britain? Statistical Self-Similarity and Fractional Dimension. Science, 156(3775), pp.636-638.
• Morris, H., 2017. 27 things you’d never know about Britain if it were not for Ordnance Survey. [online] The Telegraph. Available at: Link
• Pirsig, R., 1974. Zen and the Art of Motorcycle Maintenance: An Inquiry into Values. 40th an. ed. London: Vintage, pp. 222-225.
• Robson, C., 2011. Real World Research. 3rd ed. Oxford: Wiley, pp. 20-23(a), 16-17, 24(b).
• Weinberg, G., 1992. Software Quality Management: Volume 1: Systems Thinking. New York: Dorset House, p. 5.

Examples

• Basinger, C., 2006. Classic computer gaming, Oddware, thrifting, Tech Tales, Sims, etc [online] LGR. Available at: Link