After my recent questions I've been pondering the workings of IMAP folders some more and am still a little confused. Since an IMAP folder can (in some implementations at least) both contain other folders and messages does this mean that it can't be directly implemented as a Unix/DOS/whatever file and directory structure? In Unix/DOS a 'folder' is a directory and it can only contain files and/or further directories, it can't itself contain data. An IMAP folder on the other hand can contain both mail messages *and* more IMAP folders. Thus the implementation of IMAP must effectively be some sort of 'virtual folder' system where the folders that the user sees are not directly reflected in a hierarchical file system. Have I got this right? If so I can understand better some of the difficulties of handling this in a program, one has to both manage the mechanics of something which doesn't map onto a real file system too easiily *and* make it understandable to a user who is also used to those same file/directory hierarchies. tkrat manages the MUA problem by having a virtual folder hierarchy of its own which is kept as entries in a single file. Each entry indicates where it is in the virtual hierarchy (as the user sees it) and has a pointer to where the actual data is kept (IMAP folder, local file, POP3 mailbox, whatever). This is good because one can mix local and IMAP mailboxes and navigate among them with ease. It has the disadvantage though that the actual file names and IMAP folder names have to be carefully managed as, by default, they will all be at the same level. -- Chris Green ([EMAIL PROTECTED]) Home: [EMAIL PROTECTED] Work: [EMAIL PROTECTED] WWW: http://www.isbd.co.uk/
