I’d be keen to hear @davefarley77’s thoughts too!

If the source of many bugs is prod configuration being different to in testing, and sometimes behaviour or characteristics of external systems can differ from expectations even if the contract is valid, maybe the answer is a bit of both testing against real services and fakes?

And in the CD book it says “acceptance tests are generally end-to-end tests that run on a real working environment that is similar to production”. I think my confusion comes back to there being many different versions of what ‘end-to-end’ means!

That said, in the comments on that video he says acceptance tests are by definition running against a deployed release candidate running in the same config as will be used in production

I think that makes sense when it comes to independent deployability too, and Dave mentions using contract testing to validate the other systems’ interfaces

In Dave’s acceptance testing webinar he seems to make it clear that testing end-to-end including other systems (even those owned by other teams in the same organisation) is an antipatten as you can’t control those systems enough to test everything you need to

The timing of the TypeScript drama can’t be an accident

Great soundtrack too

I’m a sucker for look-how-clever-I-am films, what can I say

Nah try again

First impressions?

Amazing