1 public-inbox design notes
2 -------------------------
4 Challenges to running normal mailing lists
5 ------------------------------------------
7 2) bounce processing of invalid/bad email addresses
8 3) processing subscribe/unsubscribe requests
10 Issues 2) and 3) are side-stepped entirely by moving reader
11 subscriptions to git repository synchronization and Atom feeds. There's
12 no chance of faked subscription requests and no need to deal with
13 confused users who cannot unsubscribe.
15 Use existing infrastructure
16 ---------------------------
17 * public-inbox can coexist with existing mailing lists, any subscriber
18 to the existing mailing list can begin delivering messages to
21 * public-inbox uses SMTP for posting. Posting a message to a public-inbox
22 instance is no different than sending a message to any _open_ mailing
25 * Existing spam filtering on an SMTP server is also effective on
28 * readers may continue using use their choice of mail clients and
29 mailbox formats, only learning a few commands of the ssoma(1) tool
32 * Atom is a reasonable feed format for casual readers and is supported
33 by a variety of feed readers.
37 * Freedom from proprietary services, tools and APIs. Communicating with
38 developers and users of Free Software should not rely on proprietary
41 * Existing infrastrucuture, tools, and user familarity.
42 There is already a large variety of tools, clients, and email providers
43 available. There are also many resources for users to run their own
44 SMTP server on a domain they control.
46 * All public discussion mediums are affected by spam and advertising.
47 There exist several good Free Software anti-spam tools for email.
49 * Privacy is not an issue for public discussion. Public mailing list
50 archives are common and accepted by Free Software communities.
51 There is no need to ask the NSA for backups of your mail archives :)
53 * git, one of the most widely-used version control systems, includes many
54 tools for for email, including: git-format-patch(1), git-send-email(1),
55 git-am(1), git-imap-send(1). Furthermore, the development of git itself
56 is based on the git mailing list.
58 * Email is already the de-facto form of communication in many Free Software
61 * Fallback/transition to private email and other lists, in case the
62 public-inbox host becomes unavailable, users may still directly email
63 each other (or Cc: lists for related/dependent projects).
67 * git is distributed and robust while being both fast and
68 space-efficient with text data. NNTP was considered, but does not
69 support delta-compression and places no guarantees on data/transport
70 integrity. However, an NNTP gateway (read-only?) is possible.
72 * As of 2014, git is widely used and known to nearly all Free Software
73 developers. For non-developers it is packaged for all major GNU/Linux
74 and *BSD distributions. NNTP is not as widely-used nowadays.
78 * Perl 5 is widely available on modern *nix systems with good a history
79 of backwards and forward compatibility.
81 * git and SpamAssassin both use it, so it should be one less thing for
82 admins to install and waste disk space with.
86 * Stick to dependencies available in Debian main, this should make it
87 easier for potential users to install, and easier for distro
88 maintainers to pick up.
90 * A list server being turned into an SMTP spam relay and being
91 blacklisted while an admin is asleep is scary.
92 Sidestep that entirely by having clients pull.
94 * Eric has a great Maildir+inotify-based Bayes training setup
95 going back many years. Document, integrate and publicize it for
96 public-inbox usage, encouraging other admins to use it (it works
97 as long as admins read their public-inbox).
99 * Custom, difficult-for-Bayes requires custom anti-spam rules.
100 We may steal rules from the Debian listmasters:
101 svn://anonscm.debian.org/pkg-listmaster
103 * Full archives are easily distributable with git, so somebody else
104 can take over the list if we give up. Anybody may also run an SMTP
105 notifier/delivery service based on the archives.
107 * Avoids bikeshedding about web UI decisions, GUI-lovers can write their
108 own GUI-friendly interfaces (HTML or native) based on public archives.
109 Maybe one day integrated MUAs will feature built-in git protocol support!
113 * Getting users to install/run ssoma (or any new tool) is difficult.
114 The web views must be easily read/cache/mirror-able.
116 * There may also be a significant number of webmail users without
117 an MUA or feed reader; so a web view is necessary.
119 * Expose Message-ID in web views to encourage replies from drive-by
122 * Raw text endpoint allows users to write client-side JS endpoints
123 without hosting the data themselves (or on a different server).
125 What sucks about public-inbox
126 -----------------------------
127 * Lack of push notification. On the other hand, feeds seem popular.
129 * some (mostly GUI) mail clients cannot set In-Reply-To headers
130 properly without the original message.
134 Even with shallow clone, storing the history of large/busy mailing lists
135 may place much burden on subscribers and servers. However, having a
136 single (or few) refs representing the entire history of a list is good
137 for small lists since it's easier to lookup a message by Message-ID, so
138 we want to avoid splitting refs with independent histories.
140 ssoma will likely grow its own builtin ref rotation system based on
141 message count (not rotating at fixed time intervals). This would
142 split the histories and require O(n) lookup time based on Message-ID,
143 where `n' is the number of history splits.
147 Copyright 2013-2015 all contributors <meta@public-inbox.org>
148 License: AGPLv3 or later <http://www.gnu.org/licenses/agpl-3.0.txt>