Um, no one is forcing you to use slf4j or logback... ________________________________ From: logback-user-bounces@qos.ch [mailto:logback-user-bounces@qos.ch] On Behalf Of nospam.rwp@dsl.pipex.com Sent: Saturday, July 28, 2007 9:52 PM To: logback-user@qos.ch Subject: [logback-user] LOGBack and SLF4J make poor assumptions All, SLF4J and LOGBack are not backward compatible with Log4J, despite what you say: 1. Dumping FATAL looks quite stupid to me, I use FATAL when a Java Error is thrown or if my software has to quit early e.g. say it runs out of disk space or can't write vital data to a file, I only use ERROR for logging invalid data items, with WARN for duplicate data items or data items values maybe suspect, some customers require this precision so that their support staff only get called in for FATAL issues. Note Sun stupidly does not allow an ordered JRE shutdown, if you call System.exit() with non-zero value, so the CLI return value is not a viable alternative! 2. I always use the log methods with a Level because I need the flexibility to set the log level on-the-fly for some situations, the lack of log and fatal methods in SLF4J makes it unusable to me. 3. LOGBack only provides Java 1.5 source code, which is useless people who have to use Java 1.4 e.g. I have custom Log4J Appenders which were only possible due to Java 1.4 compatible source code, so no a Java 1.4 binary patch is not adequate. I know you mean well, but professionals often have to work with what is available on a system environment i.e. Java 1.4, so it is annoying that API designers don't understand this. It is often not time effective to manually edit all the generics and incompatible classes out of large code bases, in these case I have to blacklist affected APIs until all our customers upgrade to Java 1.5. Richard