LIve Blog: Immanuel Kant and the Hallucinating Tester – Anders Dinsen


(Alex) #1

Anders is starting by saying philosophy won’t hurt. Oh god he’s already talking about how we can see through our eyes but can’t actually see our eyes. Shoulda got more sleep.

In his first testing gig in a big organisation, he felt a bit tired after the first week. He’d done quite a lot of work. He was ready for the weekend. And then his test manager came and asked him what his gut feeling about the system was. His mind was a blank. He didn’t know. And that question has stayed with him.

Two things he has learned since then.

  • One week isn’t enough time to have a gut feeling about a complex system
  • That is a good question. We should expect people to have a true gut feeling.

Onto Kant. When we look to our neighbour, we think we see a friendly tester (I literally have the friendly tester next to me!!). But it could be a hallucination. When we see things around us, we see “Das Ding an sich” (the thing itself). In our mind, there is the “intuitive thing”. These two things are connected by sensing. The intuitive thing is affected by our senses. It is a hallucination, but it is affected by reality. By thinking about the intuitive thing using things that are a priori (i.e. they are born with us). The terms on which thoughts are created are shared across every single person (that’s a bit of a goosebumps moment). Fundamentally, there are four categories of a priori given terms for thinking about intuitive items. We express them using language.

(you might wanna watch the video for that slide. This talk is hard to blog…it reminds me of semantics lectures, but Anders is doing a fantastic job of it!!)

We’ve moved onto Isaac Newton. Quote: “he’s English. He’s dead now”. Laughter. He specified the first law of physics. Very empirical.

David Hume – a Scottish person (who is also dead, btw) liked this empirical nature of science, but was very worried about things like “can we prove the sun will rise tomorrow”, or “can we be certain”? He understood that what we see is represented in our minds. The same applies to proof. Our proof is in our minds. Hume was the father of scepticism.

When we look at testing, there is an actual thing, there are actual people. And there are the representations of them in our minds. What are we doing? Proving the representation? Making the case that our proof is the correct one?

If we look at software testing philosophies, there are three main categories (Anders says this is more categorical than it could be, but he doesn’t have all day). We have ISTQB, which is rather Newtonian. CDT, which is more based on scepticism. And agile, which focuses on the social.

Anders is now critiquing each philosophy. ISTQB claims empiricism, but doesn’T do much research. CDT is so focused on heuristics and scepticism that it can’t say what it knows. And agile is driven by confirmation. We say it’s ok to fail, but often it isn’t.

Anders is going to apply Kant to testing. We’re doing an experiment. We googled “test” and got around 3 billion hits. Then we googled “sex”. It got fewer hits. So testing is more important… right? There are reasons why we can explain that testing is so popular. It’s social and tied to common terms. It elevates our mind. It helps us with our gut feeling.

Oh god, this slide is about transcendental stuff. This could get sticky. Transcendental means, in this context, about knowing how to know objects (the things themselves). Knowledge is transcendental if we can reasonably know about something before we experience it (I’m gonna paraphrase with “transferable”). What we gain from testing is the transcendental relation between our a priori knowledge and the next “thing”. (My mind is thinking that this is very relevant to exploratory testing, and our expectations and our models of the world. This is probably what he wanted :)).

So, those categories – explained by Anders. We have:

  • Quantity: being or not being, and amount
  • Causality: action-reaction, relations, complicatedness
  • Quality: Beauty and complexity
  • Modality: existence and ethics

Finally, we’re talking about a philosophy of testing. Testing starts with a prioi transcendental knowledge. It critically seeks to prove the a priori. By doing that, testing establishes a transcendental relationship to reality. And that relation allows us to react rationally about “the thing itself”.

Kant can help us with testing. Testing can help us to prove things. It allows us to distrust others who stick to their a priori knowledge: they’re hallucinating. He helps us to understand why it’s difficult – relation building is not simple. It involves a lot of thinking and experimenting together. It’s also social – it helps us to understand how others think. We all share the same terms.

Anders is summing up by talking about gut feeling again: because to have this we need a prioi knowledge and the transcendental relations.

Wow, that is definitely one I need to digest over lunch. I feel like parts of my mind have been opened up and I definitely need to go through those doors at some point soon.


(Chris) #2

This is proper under-the-hood stuff that testers should be talking about! Deductive reasoning! Applicable epistemology! Humility in the face of subjective reality!


Subscribed.

Edit: DOWN WITH NAIVE REALISM!


(Ady) #3

Is this not the greatest title for a conference talk ever!

Great summaries alex, thank you for helping me share the experiences a little. :pray: :clap::clap::clap::clap::clap:

Just made me think of this!