When there are no @author tags, then there is a far smaller notion of ownership by the author(s). On one side of this, the author(s) may not appreciate changes to "their" code, and on the other side, non-authors may feel intimidated about working on code for which they are not an author. IMO, we want to _discourage_ ownership and _encourage_ everyone to work in any area of the code they want.
When there are no @author tags, we don't need a policy that says when you can add your name to a class/method as an author. I'm not even sure what our policy is, but I think we're not being consistent (other than when we create a new class/interface)
@author tags can be inaccurate. SVN has the true history of who contributed exactly what code.

The only benefit I can think of is that the @author tag does help to give some notion of who is the "expert" of the class, in case they need to be consulted. However, I don't believe this is really much of a reason, since it's far better to consult the SVN history and see who actually modified the different parts of the code. In fact, the annotated views in Fisheye even show on many of the lines the name of the last person to change it. For example, see http://fisheye.jboss.org/browse/DNA/trunk/dna-common/src/main/java/org/jboss/dna/common/i18n/I18n.java?r=120

What does everyone think about the @author tags? This isn't an official vote (we've never had one), but I would like to hear everyone's thoughts, so please reply to all and let us know what you think.

Best regards,

Randall