This topic is brought up occasionally on the RPI core. Either in writing a drop in replacement for the server that existing clients could connect to (which would require duplicating the current server protocols (including SLCP)), or in switching to a different (hopefully more adopted) chat protocol (e.g.: XMPP (Jabber), SIP,
IRC) that could be added to to support features critical to adoption as a lily replacment.
Old Features that must be preserved
# -> (11:23) From masto, to cmc:
# - Posit: the first requirement of any potintial lily replacement is that it
# - be lily.
For the purposes of this thought experiment, however, perhaps we can develop a feature list that would be "close enough."
- Must provide an upgrade path for old users. This includes those people that still use telnet only. One suggestion: provide a proxy server in front of whatever server is doing the actual CMC (--Aurelis)
- Authentication
- Accounts are not tied to anything else except having an email address
- Accounts are anonymous (the email address is private, and name can be)
- Unique naming within the entire domain (/reserve, /rename)
- Message sending
- To a single other user
- To a single other discussion
- To multiple users and/or discussions, so each recipieint knows who/what the other recipients are
- Discussions
- Persistent (Maintains membership, even if member not online at time)
- Multiple moderators
- Authorization (/permit)
- Reviews only available to members
- [optional] notification when people enter/leave discussions
- List of members available to non-members
- Existance not available to people not authorized
- Personal and discussion review
- Available only to members of a discussion
- Personal review can collect messages even when not logged on
- Flexible (parameters to narrow review)
- User attributes
- blurbs
- here/away?
- idle time
- Notifications
- Discussion creation/destruction
- User login/logout
- Ignore (both per-discussion and private)
- Server managed groups
New Features that we should shoot for because otherwise why bother.
- Rich text
- Unicode support.
- end to end encryption.
- Multiple connections by single user (screen -x)
- threads
- groups as first-class objects
- improved timezone support
Possible replacments
- Complete rewrite in non MOO engine. Aiming for a dropin replacment with extra features.
- Jabber/XMPP
- SIP
- IRC
to top