Mike Sofka's Project List.

Mike Sofka's Project List.

March 11, 2004

Priority:

WebMail Server:

  1. Webmail Upgrade Part 2: J2EE:
    1. Possible clients:
      1. JWMA (JSP+Tomcat, might not scale, needs work).
      2. Luak (Struts, not designed for email, but technology understood).
      3. Other.
    2. Requirements:
      1. J2EE.
      2. Portal integration.
      3. IMAP and POP3.
      4. Protected password.
      5. Efficient attachment handling.
      6. Personalities.
      7. Strong filtering.
      8. Virus and Spam scanning.
      9. Skins, including low-bandwidth skin.
      10. Calendar/PDA integration.

  2. Other Webmail (May not apply to Emumail 5.2.7, some are for J2EE):
    1. Virus scanning for uploaded files.
    2. Integrated spam filtering.
    3. Portal integration.
    4. Serializing folder index values (instead of re-parsing them).
    5. Correct exclusionary account locking (may not be issue without fcgi):
      1. Currently Emumail times out and assume everything will be ok.
      2. gdbm is the de facto exclusionary lock mechanism, established after emumail thinks it is locked.
      3. Instead exclusive lock USER/LOCK .
      4. Packppl, compressdb, recover should use exclusive lock.
      5. Retry, and return a ``another command has not finished'' page (or popup window) with ``Continue,'' ``retry,'' ``skip'' buttons.

    6. Personalities.

    7. Low Bandwidth/Accessible Interface (J2EE):
      1. for PDA, teletype, and readers.
      2. JavaScript free.
      3. Bitmap free.

    8. Better recognition and printing of MIME-types:

    9. Better export message and folder options:
      1. The default is unix style, which is usually wrong. Use browser HTTP_USER_AGENT string to select mac-, windows- or unix-style EOL. Add user-option.
      2. Choice of folder type and location (settable in options).
      3. Export as HTML (one file, or index+messages).
      4. Message/folder import option for UNIX mbox format, others?.

    10. TrueAttachments(tm), as needed:
      1. Replace attachments with link to file.
      2. Export attachments to RCS space (need RCS space access).
      3. Delete attachments with message option.
      4. A ``File Management'' page.

    11. Interface to RPI's PH server.

    12. Better spell checking:
      1. Whole page spell checking (see SFU's webmail).
      2. Skipping quoted sections.
      3. Better dictionary (ispell).

    13. Forward w/ header option or, a send complaint to postmaster option (that includes headers).

    14. Timeout error on login after session times out.

Web Portal Project:

  1. Quick outline of what the Web Portal Is.
  2. What should the web portal display?
  3. How will this integrate with current communication methods:
    1. Rensserv mailing list.
    2. Campus news mailing list.
    3. Mail to everybody.
    4. Memos.
    5. Documentation.
    6. Departmental newsletters.
    7. EBS/Messaging.

Mail Servers.

  1. Sendmail
    1. LDAP for mail routing and databases (Benden).
    2. SMTP-AUTH, TLS.
    3. Web page to read alias and access db and provide Gail with mail blocking information.
    4. Updating access file from /campus , via web control.

Listproc:

  1. Upgrades and enhancements:
    1. Upgrade MIMEDefang.
    2. List-owner optional spam filtering.
    3. Probe command.
    4. Graduated auto-unsubscribe.
    5. Dynamic attachment replacement.
    6. Flexible message-sizes.

  2. Listproc web interface:
    1. Support NON-RCS accounts.
    2. Switch to using ILP instead of lp2 .
    3. Listproc manager commands.
    4. Jive gateway.
    5. Rewrite in J2EE.

  3. Evaluating Post-listproc 8.2 product.
    1. mailman (http://www.list.org/) looks pretty good
    2. Also, L-Soft (http://lsoft/com/),
    3. Sympa (http://listes.cru.fr/sympa),
    4. and EZMLM (http://www.ezmlm.org),
    5. others?

Other:

  1. Math on the Web:
    1. Bring Campus TeX facilities up to standard (pdftex, ConTeXt, tex4ht, latex2html).
    2. Techexplorer/WebCT interface
    3. MathML tools.
    4. Scientific Word?

  2. Open Document Initiative:

    There are a number of good reasons for avoiding propriatary data formats.

    1. Important information can be lost as propriatary products are updated, or as companies out of business.
    2. Open document formats enfranchise more people.
    3. Open documents can be more easily itegrated with other work tools.
    4. DCMA outlaws some integration and conversion efforts.

  3. Logging host:

    Instead of saving multiple log files, which are stored, emailed, processed by various hand-spun scripts, etc., it would be better to have a single loging host which stores transaction and log information in a database. A web interface would allow authorized access from anywhere. Work on rules to display interesting transactions and trends could be shared.

  4. Campus EBS/Messaging Strategy.

    1. email: But, not everybody reads email all the time.
    2. voice-mail: Not everybody reads voice-mail.
    3. department phone tree: It works for snow days, somehow.
    4. http://www.rpi.edu/rpinfo/.
    5. The future Web Portal.
    6. Public safety going door to door.
    7. Paper bulletins in public places.
    8. Jive Frontend to EBS List?

  5. PDA integration/testing.
    1. Storing and viewing (mail) logs on-line, on the server:
      1. send to mail-log-l and archive (view w/ web).
      2. store (cooked) records and view w/ secure server CGI.
    2. Universal e-mail access:
      1. WEBMail.
      2. AFS-less universal.
    3. Ubiquitous file access (docs).
    4. Ubiquitous file editing.
    5. Ubiquitous usenet access.