Yesterday on twitter I put out automation can help you do better testing, or it can help you do poor testing faster as a bit of a vent around a problem I’ve been having the last couple days. And it might have gone no further than a sarcastic quip except for an article in the paper today talking about how all solutions have problems, and not just the normal, cheery, all problems have solutions.

Part of being a tester is being cynical. Or as I’ve pointed out a couple times, the difference between a cynic and a realist is spelling.

Its only about 1/4 of a page long, so go read it now, but here are the choice quotes.

  • Yesterday’s solutions appear on the front page of the paper as today’s problems.
  • For a solution to metamorphose into a problem it first has to work. Solutions that don’t work fade away, like Esperanto. The solutions that haunt us are the ones that function well
  • Problems are reincarnated solutions. One could say that problems get second leases on life as solutions, or solutions get second leases on life as problems
  • A solution is something that substitutes itself for the problem it has eliminated
  • There are no solutions, only replacements of one problem with another. This doesn’t mean that all problems are equal, only that all problems are problems. The best we can hope for is that our solutions will be problems we prefer to the problems they have replaced.
  • Some replacement problems are more manageable than their predecessor problems and therefore they’re easier to mistake for solutions.

So how does this relate to testing? Two things very clearly jump out at me.

First is the whole notion of automation. Testing was slow so lets put an army of machines to work! Seems like an excellent idea on the front of it, until you get tools that allow for quick, but horrible testing.

The second is information overload and the use of Metrics. Yes, the modern software development project has a lot of information to juggle and using algorithms to help navigate it is a good idea. Except when taken to the extreme and entire courses of action are determined by math and not by humans making rational tradeoffs and decisions. Just ask the quants on Wall Street how well that worked in the last decade.