> > It is a small thing in terms of code. It is *huge* in terms of thought > > and man hours. You can express the contents of the "Design Patterns" > > book in about 30 interfaces and as much default impls of those > > interfaces. However, the book doesn't really belong inside a commons > > package either. > > If the interfaces and the default implementations are *common* and don't > depend on other projects they do.
says 'commons'. Under that definition, commons would be massive. Like, half the size of all of jakarta. There is *a lot* of code that is 'common'. All of Avalon is ment to be 'common' to a lot of other projects. That's what a framework is all about, isn't it? > Where should they go else? Interfaces that > should be used in almost every Java software should of course go into a very > basic package. Maybe the next JDK would be a nice soulution. ;-))) Totally agree ;) > > Every jakarta project is able to use Avalon Framework. the > > dependency is > > captured in a single jar file, which is a stable release, > > which has been > > a stable release for a long time, which doesn't impact > > project size very > > much, etc etc etc. > > AFAIK, commons is a project that by definition doesn't depend on other > projects. is that so? I wonder why. It severely limits commons. Thinking of stuff like logging, xml, build tools, ... A project that by definition does not depend on other projects...well, by some definition you can probably call an OS kernel nondependent on other projects. But I'll draw a line right there. > If you don't want to give the interfaces fewer people will use > them. > Simple but true? yes. That's why we have OSS. I don't get it. "Everything that is reusable should be in Commons". Not a very practical goal, is it? It'd make SourceForge a lot smaller... cheers, - Leo -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
