Postby jch » Mon Nov 21, 2005 12:52 pm
Message retration has some security issues that revolve around sending a message, waiting for it to be read by some potential recipients and then restracting it when the desired level of damage has been achieved -- and leaving little trace of the evidence in the process. Tech savvy users can do the same thing without retraction, but the point here is that message retraction opens up the possibility of this kind of vindictiveness to a wider audience.
That, of course, assumes that message retraction can be made to work at all. There are several cases where implementations either fail completely or behave badly (depending on the implementation). For example:
A message appears in the summary list while you're looking at it and then disappears before you get the chance to click on it (sender has quick reactions); your interest is piqued and you then get to waste time finding out what the message was.
Retractions don't work outside of the organisation: there's no way to retract a message sent externally, although I know some Exchange users believe that there's magic involved that allows this to happen.
Messages in Scalix (and others, these days) are delivered very quickly, often with sub-second delays. This gives a recipient chance to reply to the message before the sender retracts it. You're then in the rather odd position of having one or more replies to a message that no longer exists, much to the confusion of the users that weren't in at the time.
It's quite possible to have an auto-action, server side or client side, that prints messages as the arrive. Retract that (to paraphrase Trinity).
These cases, and others like them, make me very wary of message retraction.
jch[/list]