I recently came across a module author who refuses bug reports submitted through rt.cpan.org, claiming that the system has many deficiencies, and increases the workload on module authors. The refusal is even automatic: you get an email explaining why the author dislikes RT, and the ticket is automaticallly canceled. For this person, bug reports are to be sent through regular email.
So let's think about this position.
True enough, RT has no opt-out possibility -- as soon as you have published something on CPAN, you are registered in the system, and can't decide to get out : the only options for an author are a) play the game; b) ignore tickets; c) explicitly refuse tickets as just described above. True enough, prolific authors are likely to receive many reports from RT. So for an author, it can indeed be felt as a burden.
However, for module users, it is extremely useful to have all bug reporting on one centralized platform, because this leverages collective knowledge. Even when an author is not responding at all -- and there might be many good reasons for not responding --, the system is nevertheless interesting, because you can find out if a problem was already met by somebody else, if some workarounds have already been identified, etc. When the author is responding, you can additionally know if the problem is currently taken into account or not, and sometimes there is already some patch or preliminary release to be tested.
Communicating about bugs through point-to-point plain email will just satisfy the bug reporter and the module author. Communication through a centralized bug tracking system is useful to the whole community. So, authors, please don't dismiss RT ... and if you think RT has some problems, you can always open a ticket !