There is no test for "It works, every time, always"

Notes on a twitter conversation.


  1. This is a twitter thread in chronological order to which I have attached some thoughts (mental notes I had at the time) and commentary (comments and reflections I had after re-reading the thread).
  2. The purpose of this is to document what my motivation was behind my questions & comments / tweets, and I've tried to summarize briefly.
  3. So I saw this tweet and I was puzzled...
  4. One Word To Rule Them All (HUH?) | Smart Life Tips
    One Word To Rule Them All (HUH?) | Smart Life Tips
  5. Notes: So here's the thing - who is expecting or wanting to hear that something works forever? I haven't met a product manager or VP that thinks like that. They don't intend to market their products that way. So why should development produce a product to that standard? Or really, why would a test be created to demonstrate that?
  6. That's the key point here - in what circumstances would a product be tested with the purpose of claiming it works forever. Or what is it that gets interpreted like that - and if so are there examples where I can see that happening?
  7. Hence my question about why anyone would want such a test.
  8. Notes: Then I followed on to explore where the claim about guarantees comes from, or rather if no product manager or VP is selling a product based on a guarantee to work forever, then why would the development team build anything to that standard, and why would anyone want to test for that? [To me this was sounding like a strawman statement. Anyway I wanted to try and understand this idea of a "guarantee"...]
  9. Starting assumption: The reason I'm starting from the "produced/sellable product" angle is that that has gotta be the goal that a company is working towards.
  10. Notes: There seems to be an inconsistency here - don't give guarantees - but who are asking for guarantees and why?
  11. Notes: Ok, behaving as though they want a guarantee. It still doesn't make sense if a product is not produced to that standard. Seems like an inconsistency still. This would be an inconsistency between the development of a product and my starting assumption. Let's see if this is stated like that somewhere, or if there's a reference...