Hi!
Joachim
[1]The ASF
Goals
The first goal is to have a stable but limited IMAP Server. It has to
scale well on middle-size installations. The user should be provided
with what he needs to work: A centralized inbox and the ability to
sort messages into self defined folders.
This goal should be reached fast with just the features needed to
achieve it That means we should analyze clients and notice what they
are doing and how they react on limited features.
Release Early, Release Often.
The community feedback is the only chance for such a project to
survive.
Apart from that it is also the right time to talk about visions.
IMAP has a lot of new requirements which means that existing design
of
James message and user storage API backends have to be extensively
refactored/rewritten. What ever we do now may be the base of future
development. Design limitations we introduce today could be hard to
overcome tomorrow.
Having a structured, ordered and valued roadmap of what ever is
desireable/imaginable helps on making decisions.
Being pragmatic may mean discarding some goals, but we should at
least
be aware of that.
Roadmap
Milestone 0 : Preparations
1. Define goals
2. Define a roadmap
3. Start a glossary
4. Collect resources
5. Design an backend API that is in line with the features desired
in
the roadmap
The API could consist of a bunch of interfaces. They don't need to be
complete but should allow for future extension.
Milestone 1 : The Basics
* Stable
* JDBC based ImapMailboxRepository
* Internal use of namespaces
* Full uid support
* Support for all basic operations for at least one client (Mozilla
Thunderbird)
Supported commands: Partial supported: Not supported:
CAPABILITY FETCH EXAMINE
NOOP Not supported: SUBSCRIBE
LOGOUT FETCH (BODY[<section>]<<partial>>) UNSUBSCRIBE
COPY FETCH (BODYSTRUCTURE) LSUB
LOGIN FETCH (ENVELOPE) CHECK
SELECT FETCH (FULL) CLOSE
CREATE EXPUNGE
DELETE SEARCH
STATUS X
APPEND AUTHENTICATE
STORE
Milestone 2 : Full Compliance
* Nearly full compliance to IMAP4rev1 RFC2060
* SUBSCRIBE will still have no effect
* Support for all non-buggy IMAP4rev1 compliant clients
Milestone 3 : Compatibility
* Maybe/hopefully not needed
* Maybe (partial) IMAP4 compliance
* Workarounds for buggy wannabe-IMAP4rev1-compliant clients ;-)
* [2]Postel's Law "be conservative in what you send, be liberal in
what you receive"
Milestone 4 : Shared Mailboxes
* Full compliance to IMAP4rev1 RFC2060
* Subscriptions
* Basic ACL support, to be controlled by admin
* Shared Mailboxes
Milestone 5 : Fancy Groupware
* Quota
* Sieve integration
* User-defined ACL
* User-defined Flags (Keywords)
Milestone 6 : Cluster
IMAP is quite resource intensive. Mailboxes might reach the size of a
few GiB. Clients are not required to do any caching. Users might
search and wade through large message archives from day to day.
* Large enterprise installations
* Many imap servers to many storage servers
* Automatic fallback to backup stores
* Fine-grained performance audit
Milestone 7 : Imagine
* IMAP mirroring
* NNTP-Integration
* Push IMAP
_________________________________________________________________
Copyright � 1999-2006, The Apache Software Foundation
Verweise
1. http://www.apache.org/
2. http://en.wikipedia.org/wiki/Postel's_Law
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
