0 Replies Latest reply on Apr 26, 2005 5:37 PM by Andrew Oliver

    Scaling back JBoss Mail Server 1.0 and planning a 2.0

    Andrew Oliver Master

      So guys, we can't let this linger until the end of the year the way I'd planned. Lets reprioritize and nail what we need for a marketable 1.0 mail server and then lets make this the cutting board for milestone releases.

      * Stabilize our current stuff - (Mailboxes and blobs, etc)
      * Security - Basic association of security (roles association for Create, Delete for mailboxes, role association for SEND) should work with jboss login modules. Do this in JAAS and its done.
      * Virtual Hosts - acoliver@jboss.org != acoliver@federation.jboss.org
      * IMAP - basic IMAP for mail server support
      * traditional spam filtering support - just someone needs to write the mail listener
      * JCA - needs to be JCA adapter for sending/receiving mail
      * WebMail - need to have at least a disconnected traditional webmail (later we should have a "direct connected"
      * Mail Lists - including adminsitration features
      * Mail Forwarding - let users forward their mail elsewhere
      * Mail routing - set this up as a middle tier mail server
      * "clusterable" - in a "won't break" in a cluster sense of the word (but nothing specific)
      * Load test - should handle load.
      * Installation - easy installation
      * Administration - easy administration, possibly work with JBoss Admin Console folks

      2.0:

      * Mail API standardization - meaning a more sensible "user" api for writing mail listeners
      * REST support
      * Exchange integration
      * Calendar
      * Mail Scripting
      * NIO and other non 1.4 pie in the sky stuff
      * Exchange class adminstration
      * High end clustering/caching
      * Bigger spam tools (fingerprints, etc)
      * ... etc.

      I'm going to start aiming the milestones at that in 1.0 instead of the "big picture" that will be 2.0. My original opinion was that we wouldn't be able to really be marketable until we had the whole thing; however, with interest building and usage already starting I want to be able to bring people into professional development sooner rather than later. Building a bigger base really means having a release that we're comfortable "blessing" as production. Then we deal with real users, real support and pain.

      Thoughts? Questions?