On 2/9/03 19:33, Lee Larson wrote >On Sunday, February 9, 2003, at 03:50 PM, Bill Rising was confused >about: > >> If I connect to my iTools account, I can create folders as expected and >> move mail messages around as expected, for instance. >> >> Make a new folder called Foo >> Put a couple of email messages in it. >> Decide it should have a subfolder. >> Make a subfolder of Foo called Bar >> Put a couple of messages in it. > >I don't think this behavior is correct. Doesn't that leave the location >of the original messages within Foo undefined?
Hmm. When I read the specs, it says that there are two allowable behaviors for specifying mailboxes: *requiring* the trailing slash to state that the user is referring to a mailbox, and *ignoring* the trailing slash. If the trailing slash is required (and if mailboxes got invisible extensions, like .imapmbox, like they do in ~/Library/Mail), the there would be no naming conflict between Foo/ and Foo when talking to the server. Of course, the server would have to have some sort of extension on the end of the mailbox name (like .imapmbox). So... it seems that if the mail server understood extensions for mailbox names, there could be both a Foo mailbox and a Foo folder holding other mailboxes. This would take care of the above naming conflict. Perhaps there is some way to get the mail servers to run folders the same way as a file system. Otherwise, IMAP is pretty clumsy as a method for keeping track of info, because it requires the user to know from the outset which containers (folders) will be end nodes (mailboxes) and which will allow branches (be directories). So... if I decide that I need to split a topic because of having to many messages, I have to make another directory with subdirectories, and then move all the mail messages over. Ugh. Somehow it seems hard to believe that something as widely used as IMAP would have this rather basic limitation. Bill | The next meeting of the Louisville Computer Society will | be February 25. The LCS Web page is <http://www.kymac.org>.
