Summary

The aim of the project is to create a user-friendly, extensible, and scalable search package. The project team consists of Dirk Gomez (co-lead), Deirdre Kane (co-lead), and Dave Bauer. MIT Sloan is the lead .LRN institution for this project.

There currently is no working search solution for .LRN. The document below describes how to implement a search package for .LRN for Oracle 8i, which is the database system in use in MIT Sloan School. The described approach should be portable to PostgreSQL.

Packages/portlets used by Sloan

A User may have these portlets:

• News
• Useful Links (custom portlet)
• Links
• Day Summary
• Full Calendar
• Documents
• Surveys
• Forums
• Frequently Asked Questions

A class may have these portlets.

• Class Info
• News
• Frequently Asked Questions
• Forums
• Syllabus
• Staff List
• Schedule
• Subgroups
• Survey
• Full Calendar
• Documents
• Calendar
• Homework
• Assignments
• Lectures Notes and Slides
• Handouts and Study Materials
• Trading Game

A communitiy may have these portlets:

• News
• Frequently Asked Questions
• Survey
• Forums
• Subgroups
• Documents
• Members
• Full Calendar
• Community Info
• Calendar
• Schedule
• Syllabus
• Trading Game

Portlets/packages not indexed

• Trading Game
• bookmarks
• curriculum
• news-aggregator
• blogger

Work on packages

To be reviewed

• faq
• survey
• static

First draft programmed

• calendar
• curriculum
• dotlrn

Pending

• File-storage
• news

Per-Package docs and comments

FAQ

Index questions and answers and index FAQs separately.

Forums

Index messages and forum title charter separately.

Content-Repository

See this Forum thread

Open Questions and Tasks

Should we use triggers for indexing content? Contrast portability, stability, plugin-ability.

Read up on doc_highlight

Maybe store the community_id in the intermedia content?