Hello Niko Wittenbeck,

Thanks for the report. You're right, we should accept booleans as well as strings.

However, this is easily worked around, and affects not only boolean types, but all other primitive types, and maybe also some more complex configuration properties (we could imagine accepting object instances where class names are usually expected, for instance, or in some cases we could accept lists of Strings).
Moreover, this can be fixed more easily and cleanly in Search 6, where a more advanced property parsing infrastructure was introduced (see here and there, with an example there.

So I'm going to push that to Search 6 for now.