Like all things DDD, there are no hard-and-fast rules here but merely
design trade-offs. Does the explanation provided on the Characterization
page (
https://java.net/projects/cargotracker/pages/Characterization)
under the "Domain Services" section help much? If not we can try to
clarify again here.
Basically in this case the domain service implementation is so
infrastructure heavy that the implementation resides in the
infrastructure layer while the domain logic signature (interface) is
still in the domain layer. If it were purely a local domain service
implementation it probably would have belonged strictly in the domain
layer (this is really the most common case for most domain services
since they tend to be quite business logic heavy rather than
infrastructure heavy).
On 10/2/2015 10:13 AM, dehms_at_gmx.de wrote:
> I would like to discuss the question what kind of DDD service the
> RoutingService is. I would call it an infrastructure service because
> its implementation is located in the infrastructure layer. But in one
> of Reza’s talks he said it is a domain service. So maybe I missed
> something?!