From owner-freebsd-hackers Fri Mar 31 09:48:51 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id JAA29030 for hackers-outgoing; Fri, 31 Mar 1995 09:48:51 -0800 Received: from aries.ibms.sinica.edu.tw ([140.109.40.248]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id JAA29023 for ; Fri, 31 Mar 1995 09:48:47 -0800 Received: (from taob@localhost) by aries.ibms.sinica.edu.tw (8.6.11/8.6.9) id BAA01594; Sat, 1 Apr 1995 01:48:10 +0800 Date: Sat, 1 Apr 1995 01:48:09 +0800 (CST) From: Brian Tao To: Terry Lambert cc: freebsd-hackers@freefall.cdrom.com Subject: Re: Mail... In-Reply-To: <9503291957.AA20552@cs.weber.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: hackers-owner@FreeBSD.org Precedence: bulk On Wed, 29 Mar 1995, Terry Lambert wrote: > > Mail really wants a record oriented file system, or as you suggest, one > file system object per message. Precisely. Since we are all used to seeing hierarchical filesystems, this model can be applied to a mail spool rather well. Each individual user's mailbox is represented as a directory structure, where the "inodes" make up the mailbox index. Each message is represented as an individual file inside that directory. If you want to get fancy, you can map each message to a subdirectory, with the headers, message body and MIME attachments as files. Large organizations could then build up trees of mail filesystems and instruct mail readers to descend a hierarchy of directories the same way you would trace from a top-level domain down to the invidiual host and then to the user. But this is quickly digressing from the original matter. I only suggested a mail filesystem half-seriously, but it is an intriguing problem. FreeBSD should be the first to do it. :) -- Brian ("Though this be madness, yet there is method in't") Tao taob@gate.sinica.edu.tw <-- work ........ play --> taob@io.org