Klaus Alfert

The kind of relevant statistics depends usually on the properties or the system under test. In your example, I would be interested in the length of the list. ```elixir check...

I prefer a functional approach, but without macros, if feasible. If you look at the F# examples (which are quite similar to propcheck and Proper), than we could enrich the...

This would not work, if you want to use several aggregations at the same time, which is usually the case when you classify the generated data e.g. into buckets like...

@josevalim `assert` raises, that is very true. But if not, than it returns `true`. In other frameworks the result of a property is usually simply a boolean. In `StreamData` you...

@whatyouhide Properties are by their very nature boolean statements. But the `assert` approach is very pragmatic and helpful, in particular for the error messages. Let's stick to it. Did I...

@whatyouhide Could you please rewrite my last example in your way? I am not sure how I would have to formulate it properly. Perhaps I am thinking in a complete...

Ok, so if the property fails we have also the stats for failing property run. This is good! I think that `aggregate` and `classify` are macros, otherwise they have no...

Correct, but when assert raises, the following commands are not executed. So, unless you reorder the statements in the property as part of the implementation , their order is significant....

@josevalim sure: property testing of stateful systems in Quickcheck, Proper, and others work with a state machine as the abstraction of the system under test (SUT) and a command generator,...

@josevalim Depends on what kind of API you are thinking about. For the property testing, I think about following either QuickCheck or PropEr closely - this is work in progress...