[vertex-four]

What reasonable default behaviour? If you're writing a console application, dumping to stderr is probably not a good idea - writing random log files that the end-user/admin has no idea about is probably not a good idea - shutting up and doing nothing is probably not a good idea (Java code has a tendency to use logging to warn about non-fatal configuration errors, etc). And let's not get started on what verbosity to set by default.

[btilly]

A good default is to write to STDERR if the severity level is above some threshold, otherwise do nothing.

That way the console can tell you about configuration errors, and by default the library will log nothing. If the library author considers it important to log somewhere, the library author is responsible for putting both an understandable warning up, and writing documentation.

[marcosdumay]

If you are creating a library? You shouldn't log at all. Return errors or throw them - do not log.

If you are creating a console application? Stderr is the standard way to log errors. Let the user decide if where it points to.

If you are creating a daemon, set the log destiny at the launch script. The application can either log to stderr anyway and let the script handle everything, or offer a parameter for setting the log channel.

The Java community fixation on logging is really not deserved. It's more caused by bad error handling by their main web servers than by any fundamental problem.

[empthought]

Logging is useful for more than just error conditions; how does that square with your advice to library authors?

[marcosdumay]

If it's not an error condition, you return it.

Libraries must not do any covert IO (in fact, any covert anything). A logging library is the only one that should log, and not their own messages.

They may write into some main application supplied collector, if it is not global, or if it is clearly handled as a message passing API.

The Java way is simply bad.

[empthought]

Maybe I'm misunderstanding your point -- by "do not log," do you mean do not call logging statements in a logging library's API, or simply do not force a concrete implementation of a logging library? I understand and agree with the second one completely, but the first one seems unrealistic.

I can't see how the Apache httpclient library could sanely "return" the different state transitions and pieces of data that are logged in its wire-level logging, but this is certainly a case where the logging statements are welcome.

[marcosdumay]

Yes, if you created a framework, that will call your code, yes, it can log. It can also provide a logger implementation, or request one from you, whatever it deems better, and your code must comply.

But a library, that your code will call, must not log. Ever. That includes any portion of a framework that you have to call, but it using services upper on the stack can make this rule hard to see. It must not include a logger implementation, and not call any global logger supplied from you.

It seems we are arguing over the definition of those two cases.

[empthought]

So in the case of Apache HTTPClient (which I would call a library), are you saying that it is a design flaw that it provides logging statements of its internals, when provided the proper configuration/logging implementation?

If so, I would have to disagree based on pragmatic grounds, unless you have another credible way to acquire the information present in http wire logging or connection pool events, for example.