What I'm saying is that the term ``whitelist'' doesn't imply ``release'', therefore releasing after whitelisting the address is non-intuitive behavior.
Now that we're aware of the differences in terminology/intuitions, we can clearly define (or redifine) the meaning of Whitelist for everyone, so that intuition is not needed any more by anyone who has 'rtfm'.
It's so non-intuitive that I wasted several days on Jeff Ross' bug report because I never considered this was the case.
As for the confusion with Jiff Ross's bug, when PENDING_WHITELIST_RELEASE is added it won't matter what the default setting is, really, as we'll alway have to ask to what (if anything) they have set the value of this variable every time we get a related support request.
As for what the default should be set to, I suggest that we put it to a poll on the User list. Whatever the users of tmda-pending and tmda-cgi think "Whitelist" should mean should be the default, no matter what you or I think it should be, so that it is easier for those end-users to configure TMDA.
In other words, if we pick the majority's interpretation, knowledgable users who read documentation will properly know if "Whitelist" means append+release or just append, and those who can't be bothered will have the default action being what they expect, most of the time.
It may be useful to note officially that there will always be a difference in my mind between what the innards of TMDA mean when they say "Whitelist" and what the interfaces to the queue (tmda-pending and tmda-cgi) mean when they say "Whitelist", as these *are* different things.
In TMDA we are dealing with email addresses, no actual email involved yet. Here it means literally "add an address to a list that is called whitelist so that from now on I get all this sender's mail".
In queue interfaces, people are dealing with messages and not addresses, so whitelisting implies that you are accepting the sender and the sender's message.
I think that's the big difference (email versus email&message) we're trying to reconcile now.
-- Aside -- This is looking more and more like the debate about "Salvation through faith alone" ;)
So, if you want a behavior that does whitelist + release, we need to find another term for it so that it doesn't trip others up like it did me.
One I've thought of so far is "Confirm". Of course this may be a little confusing since that's what _other_ people do when they respond to a challenge. Would it be enough distinction to say that when OTHER people confirm it implies that TMDA check CONFIRM_APPEND to see where to put the address, and when _I_ confirm a message in a queue it implies it uses PENDING_WHITELIST_APPEND?
Another idea is to use "Accept" since I want to not just release this email from the queue but also accept any further email from the sender.
If choosing a new name is the direction we go, I don't think we'll need PENDING_WHITELIST_RELEASE at all, as the new term can mean "Release the email and append the sender to either PENDING_WHITELIST_APPEND or PENDING_RELEASE_APPEND, whichever is defined".
I still vote for keeping the word "Whitelist" and redefining its meaning
in the context of queue interfaces, as I believe it is more intuitive to think of the message and the sender as inseperable in that context, but I'd really like to get a broader opinion base on what the default of PENDING_WHITELIST_RELEASE should be.
-- Jim Ramsay
_________________________________________________ tmda-workers mailing list ([EMAIL PROTECTED]) http://tmda.net/lists/listinfo/tmda-workers
