distributedlife

Today I was asked to raise a defect because the background colour in the header was different to the background colour in the footer. I politely pointed out that the colour variation was more than likely caused by the LCD monitor and the viewing angle. “Definitely not!” was the reply and I sheepishly agreed. I should be berated for even considering such a thing.

To be a thousand percent sure I went back to my desk and checked. Wouldn’t you know it. Identical colour codes. What could cause the mystery of the colour variation?

Here is the Wikipedia entry for Liquid Crystal Display

Here is a quote from the drawbacks section:

LCD panels using TN tend to have a limited viewing angle relative to CRT and plasma displays. This reduces the number of people able to conveniently view the same image – laptop screens are a prime example. Usually when looking below the screen, it gets much darker; looking from above makes it look lighter. Many panels which are based on the IPS, MVA, or PVA panels have much improved viewing angles; typically the color only gets a little brighter when viewing at extreme angles.

This brings me to the title of this post. Testing is in the eye of the beholder. Which is wrong. Testing is, generally speaking, a quantitative process. The eye of beholder metaphor is about qualitative judgement. In this example we solved a qualitative judgement using a quantitative approach. We check the CSS file and proved that although our eyes perceived a colour variation the colour did indeed match the specification.

This now concludes my argument on why testing is a specialist skill like every other facet of the software development life cycle.

note: I won’t go into when qualitative testing is appropriate. I’ll save that for another post. It is a justifiable approach, just not in this scenario.

2nd note: I’m sure not if I added enough sarcasm to the first part of this post.