Hello Thomas, On Tuesday, September 9, 2003 at 7:58:20 PM you [TF] wrote (at least in part):
TF> My point is that the user, while not RFC-conform, has expressed a TF> clear intention who he wants to send the email to. And that's the point: he did not. If he would have done TB! wouldn't have to "guess". TF> TB concatenates in a way that sends the message to another address TF> than intended (mutt doing the same is an irrelevant information). No. It's a relevant information in so far as mutt is an old and established MUA known to be very close to RFCs. It's also known to really seldom violate RFCs. And if even mutt "assumes" the same as TB! does chances are high it's something in the users input that is *not* unambiguously. TF> Your point is that the user violated the RFCs. No. User did not send the message therefore he (she) did not violate RFC. User just gave input that's not RFC conform and indirectly asks The Bat! to make the best of it. I see your point an absolute novice can't know a MUA needs a "delimiting character" to be able to distinguish an address from a "pure name": the '<'. But there's nothing a MUA can do about this, except offering two input fields: "Name" and "E-Mail-Address". Create a new mail, type in the address field: Thomas [EMAIL PROTECTED] *YOUR* intelligence tells you the latter is "the address" (and therefore needs to be embraced in angle brackets). But this intelligence can't be put into Delphi-(C/C++/whatever) code that simple. Imagine this address line given by the user: Thomas Thomas What I "meant" was: Thomas <Thomas> absolutely valid as an e-mail-address. You see the problem? Without a unique and unambiguously delimiting character The Bat! cannot safely and securely recognized what the user "really meant". One can argue about "the '@' is a unique identifier to recognize the e-mail-address". But this leads to another problem: [EMAIL PROTECTED] [EMAIL PROTECTED] What about this "user input"? Could be two addressees. Or could be '"[EMAIL PROTECTED]" <[EMAIL PROTECTED]>'" (e.g. in opposite to '"[EMAIL PROTECTED]" <[EMAIL PROTECTED]>'). TF> How will we find a solution? There are two possibilities: TF> 1.) Force very user to read the RFCs. TF> 2.) Send the message to where he declared it should go. TF> Which do you think is the more sensible? We both know #1 is not realizable. But your second suggestions lacks a connection to reality. The point is: what we're discussing about is a scenario where the users did *NOT* declare where the message should go to; declare in a form an Internet message needs declaration. So I'd suggest somebody creates a wishlist item that asks to change the behavior to "split at space character and if in doubt only use the last chunk that could be syntactically interpreted as valid e-mail-address". -- Regards Peter Palmreuther (The Bat! v2.00.6 on Windows XP 5.1 Build 2600 Service Pack 1) Anyone can hold the helm when the sea is calm. -- Publilius Syrus ________________________________________________________ Current beta is 2.00.06 | "Using TBBETA" information: http://www.silverstones.com/thebat/TBUDLInfo.html
