[asim]

Probably not but users tell me when something isn't working. I spent my entire career in this sort of development model where 80% was good enough, put it into production and then moved on. When something breaks, fix it, when something is really complex test it, when the logic is critical mock it.

[robertlagrant]

This is likely to be a core issue for most people. In reality it isn't "if it breaks, fix it", it's "if we notice it breaks, and we can remember what it should do, we can probably fix it". Tests allow you to spot when you break things, and also encode what they should do.

[adrianmsmith]

Tests also cost money to write and to maintain. It's a tradeoff. Personally I like to work at places where the tradeoff falls in favour of writing tests. But I understand there are some businesses that just don't have a lot of money and where the cost of things breaking isn't that high, and then the right decision might be to skip tests. And one might argue if the customers don't notice it's broken, does it really matter?

[robertlagrant]

I agree they cost something, although I think they give back far more than they take. In this case though, I would say that not having them in place means people won't adopt these proffered microservices, as everyone will have to individually implement tests around them.

[throw827474737]

That must have been then completely early startups, web-focused? Even there cannot imagine how this could be.. just wow.