http://bugzilla.slf4j.org/show_bug.cgi?id=257 --- Comment #2 from Irv Rabin <ipr1118@gmail.com> 2012-05-09 20:47:43 CEST --- Thanks Joachim, I expected a response like that. :) Well, logging is not only "to collect information to help developers fix the problem". That is debugging. Only logging messages with level DEBUG fit into your definition. Well, perhaps TRACE to a certain degree. Logging is to collect data about a running process so the latter can be analyzed later. For example, a big commercial billing application may want to have a log where the whole billing process can be traced and reconstructed. The application may have thousand of classes. Messages relevant to one billing cycle that logically belong to the same logger could be produced by any of the classes comprising the application. The logger could be defined for a whole subsystem (such as Billing). Then, as the application can be divided into subcomponents, the logging could be distributed into subloggers (Billing.Notification, Billing.Payment, etc), so the dots would still show hierarchy and configuration with inheritance of settings could be honored. But whenever the system is composed of tens of thousands of classes, using class names as logger name is simply insane. Like a motorboat driver has a very vague idea how to drive an aircraft carrier ship, in the same way recommendations that are sound for small Proof-Of-Concept projects, are inapplicable to big commercial applications. That's why I believe that recommendation of using class names as logger names should be limited to small pilot projects, where the developer is the primary user of his own code. -- Configure bugmail: http://bugzilla.slf4j.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug.