Learning/Event Service
Minutes: 9 November 1998 6:30 Smith Hall 100
PRESENT
  • Rudy Setiawan
  • Jon Hsieh
  • James Lampe
  • Yun-Ching Lee
  • Wing Ling Leung
  • Jon Wildstrom
  • Andy Zimdars
  • Eric Stein
STATUS
  • James:
    • Leader for the implementation phase
  • Rudy:
    • Will meet with Network team leader after Tuesday's class
  • Jon H:
    • Will meet with Database team
    • Documentation meetings for the last few weeks were cancelled
  • Jon W:
    • Will miss next week's meeting, will try to catch up by reading minutes
  • Andy:
    • Architecture meeting result:  will have real data for demo
  • Wing:
    • SDD presentation
    • Evaluated the statistical modeling package, need to separate algorithm from GUI
  • Yun-Ching
    • SDD presentation
DISCUSSION
  • Event Services
    • Database's concept has more direct subsystem interaction
    • Resolution:  Establish Event Services API and force everybody to use it
    • Preliminary API:
      • StartUp()
      • ShutDown()
      • Publish()
      • Subscribe()
      • CreateChannel() (maybe?)
    • Also need to find out how much of Network's events overlap
  • Data Miner
    • Limitation of Rudy's API:
      • Input may be continuous or discrete values (okay)
      • Output can only be discrete (may be a problem)
    • Need to finalize on DataMiner's input/output needs
    • Need a training data set for implementation/demo
    • Scenario 4 (M-class):
      • frequency of data access
      • data table type requested
      • connection speed/dealer's hard disk space
    • Scenario 5 (minize connection cost):
      • timestamp of request and delivery
      • network ping time for bandwidth
      • COMET server load (DB hits or CPU load?)
    • Scenario 3 (morning congestion):
      • logon/logoff time
    • Scenario 6 (detached operation):
      • cache problem, not applicable to learning
    • On low bandwidth systems, request the dealer admin to leave the system connected 1 night out of a week/month for update
    • Split up into teams and formalize scenario needs for Data Miner
  • Preference Panel:
    • Need these values:
      • Radio buttons:
        • Smart update - no prompt, update automatically
        • Manual update - no prompt, manually initiated updates
        • Confirmed update - prompt with timeout
      • Integer:  timeout for confirmed update
      • Boolean:  download after timeout
    • Send these requirements to UI team immediately, and more as they come up.
ACTION ITEMS
  • All:
    • Put the team in your zephyr .anyone file so we can zephyr questions
    • Help Jon W and Yun-Ching with Java issues
    • Don't forget Homework 3
  • Eric:
    • Find out Helmut's version of SDD -- may not be the latest
    • Check out the disappearing notification problem with bboard
  • Rudy:
    • Report results of meeting with Network group
    • Release Java source of DataMiner interface
  • Jon W:
    • Send Preference Panel requirements to UI team
    • Implement Scheduler
  • Yun-Ching:
    • Draft Event Services API
    • Identify Learning in other scenarios (not 3, 4, or 5)
  • Rudy, James, Andy and Wing:
    • Get together and interface Data Miner with learning algorithms
  • James and Rudy:
    • Identify DataMiner input/output requirements in Scenario 3
  • Wing and Jon W:
    • Ditto, Scenario 4
  • Andy and Jon H
    • Ditto, Scenario 5
CRITIQUE
  • Well run meeting overall
  • Need to start coding very soon

This page is hosted by the Chair for Applied Software Engineering of the Technische Universität München.
Imprint (Impressum)