[Alpine-info] the TAB key

[Eduardo Chappa]

On Sun, 27 Jan 2008, Mark Crispin wrote:


:) On Sun, 27 Jan 2008, Eduardo Chappa wrote:

:) > Nevertheless, \Marked means nothing, even if most people use it to 

:) > mean *it's a fact that there are recent messages*, it might not mean 

:) > that.

:) 

:) That is nonetheless the intended purpose of \Marked.  Something changed 

:) in the mailbox that has not been read; and that is almost certainly a 

:) recent message.


No, that is not what it says. I know you wrote it, but you did not write 
what you meant to say. If I say "I will do my homework by 7 pm" is way 
different to "I will probably do my homework by 7 pm". One of the 
definitions uses the work "probably", the other uses a "certainty". In 
the language of logic, one can say that the property of being \Unmarked 
is a statement, the other is not.


:) > The RFC is quite ambiguous about its meaning and for all I read, you 

:) > can return \Marked and \Unmarked in the same response.

:) 

:) You misread, and this may be due to a language barrier.  \Marked and 

:) \Unmarked are mutually exclusive; \Unmarked is a definite statement.


Show me how i misread. I do not see it. \Marked is "probably", \Unmarked 
is "certainty", the negation of "certainty" is "certainty", not 
"probably".

If my argument is based on a language barrier, I would like you to show me 
the word I am misintepreting. To the word "probably" is not the same as 
"it does", and it is the basis of my reasoning.


:) There is, however, a third state: neither \Marked nor \Unmarked.  In 

:) that case, the server is not saying anything either way.  That is why 

:) there are two flags; it is a three-way state.  If it was a binary state 

:) there would only need to be a single flag.


I agree that there is a third state, what I claim is that due to the 
ambiguity, both answers are allowed, so there are four.


:) > So the contract of the TAB key is the existence of RECENT messages, 

:) > and Alpine uses the LIST command to check for that. That's not 

:) > correct. That's the bug.

:) 

:) We disagree.


fair enough that you state your position, but your claims still do not 
disprove mine. LIST is not command to check for RECENT messages, it is a 
command to check failure of existence of recent messages. I think you need 
to address in the RFC what is the meaning of "probably" in the \Marked 
response. Two servers may report differently on the same mailbox just 
because of ambiguity in this definition.


:) > I understand the advantages of using the LIST command, but until 

:) > \Marked really means there are new messages in the folder, I do not 

:) > believe Alpine should be using it, just to please those who like 

:) > their software gain speed for the price of doing the right thing is 

:) > not right.

:) 

:) We disagree.


Which claim do you disagree with? all? any in particular?

-- 
Eduardo
http://staff.washington.edu/chappa/alpine/