Transcript

Mobile Computing at Acadia University

Tomasz Müldner

Jodrey School of Computer Science

Acadia University, Nova Scotia, Canada

October 13, 1998

Contents

• Electronic campus at Acadia University• Related software projects:

– Automated Courseware Management Environment– Shared Workspace– Distributed Marking System– Jersey

• Other related commercial software• Mobile agents vision

Contents

• Electronic campus at Acadia University• Related software projects:

– Automated Courseware Management Environment– Shared Workspace– Distributed Marking System– Jersey

• Other related commercial software• Mobile agents vision

Acadia University

Acadia University:Overview

• Acadia University was born in 1838

• It is small (< 4,000 students)

• Concentrates on undergraduate programs

• In 1996, partnership with IBM formed

• All faculty members and students get an IBM laptop

Acadia University:Electronic Classrooms

• All classrooms will have:– instructor’s console– network connection for every

student

• Two kinds of classrooms: lecture and studio

Acadia University:Electronic Classrooms

Contents

• Electronic campus at Acadia University• Related software projects:

– Automated Courseware Management Environment– Shared Workspace– Distributed Marking System– Jersey

• Other related commercial software• Mobile agents vision

Related Software:ACME

ACME (AITT) provides a single place for:– course-related information:

course outline office hours

– discussion groups– on line testing– upload and download of notes

Related Software:ACME

Related Software:ACME

Related Software:ACME, Sample Experience

Related Software:Implementation of ACME

• Server-based, CGI scripts written in Perl

• Dynamic HTML pages

• JavaScript for visual enhancements

• Standard file system for persistence

Related Software:Discussion of ACME

• Popular: used in many courses

• Does not require any installation (Web browser)

• Tends to overload a Web server

• Support for pull, no support for push

• Does not provide collaborative facilities

• Does not automate various tasks

• No support for disconnected operations

Contents

• Electronic campus at Acadia University• Related software projects:

– Automated Courseware Management Environment– Shared Workspace– Distributed Marking System– Jersey

• Other related commercial software• Mobile agents vision

Related Software:Shared Workspace, SW

Provides support for:

• asynchronous exchange of information

• synchronous collaboration in electronic (virtual) classrooms

Related Software:SW: Exchange of Information

• Asynchronous sharing of homogenous, centralized information systems, IS

• Each IS consists of classifications and information

• Every SW can fetch and provide

Related Software:SW: Example Repository

Related Software:SW: Collaboration

• Support for collaborative editing with a single controller and multiple ghosts

• Various floor control policies

• Chat room for “raising a hand”

Related Software:SW: Collaboration

Related Software:SW: Implementation

• Every SW is a server and a client

• Client/Server implementation using Java and RMI

• JSDT implementation of shared editing

Related Software:SW: Discussion

• Limited support for collaborative editing

• Inefficient updates of ghosts

• No support for filtering, locating, automatic organization, or alerting

• Each IS is centralized rather than distributed

• No support for disconnected operations

Contents

• Electronic campus at Acadia University• Related software projects:

– Automated Courseware Management Environment– Shared Workspace– Distributed Marking System– Jersey

• Other related commercial software• Mobile agents vision

Related Software: DMS Functionality

• Electronic maintenance of assignments

• Management of distributed information resources:– assignment descriptions– assignment solutions– marks

Related Software: DMS Functionality

• Instructor submits assignment descriptions

• Student downloads assignments and uploads solutions

• Marker downloads solutions, and marks them off-line using specialized marking software; then uploads results

• Student gets a notification

Related Software: DMS Student

Related Software: DMS Marker

Related Software: DMS Implementation

• DMS consists of a central server and set of “users”: Students, Markers, Instructors

• A user may need to download its application and run it locally. DMS will package the application and deploy it to the user

Related Software: DMS Implementation

•Shared resources and the software that controls access to them resides at a server site

•Java and CORBA are used (for language independence and speed)

•Java Applets are used for the user-side client (for (re)deployment)

Related Software: DMS Discussion

• Centralized server creates a bottleneck

• Very limited support for push

• For marking each assignment, a marker has to develop an ad-hoc marking strategy

Contents

• Electronic campus at Acadia University• Related software projects:

– Automated Courseware Management Environment– Shared Workspace– Distributed Marking System– Jersey

• Other related commercial software• Mobile agents vision

Related Software:Jersey (Tomek, Giles, Nicholl)

MOO/MUD (Multi-User-Dialogs) based systems real world emulation (spatial)

–Users, places, objects–Navigation, communication–Creating, modifying, transporting objects–Run-time customizability, expendability

Related Software:Jersey: Implementation

• Client/server written in VisualAge Smalltalk• Communication with the user by Smalltalk messages (extended with the list of messages)• “Agents” used to automate user queries, automatically capture events, etc.

Related Software:Jersey: Discussion

• Spatial rather than functional organization• Traditional client/server• Agents are really daemons

Contents

• Electronic campus at Acadia University• Related software projects:

– Automated Courseware Management Environment– Shared Workspace– Distributed Marking System– Jersey

• Other related commercial software• Mobile agents vision

Related commercial software

• Outlook Express:– Scheduling– Email Rules

Contents

• Electronic campus at Acadia University• Related software projects:

– Automated Courseware Management Environment– Shared Workspace– Distributed Marking System– Jersey

• Other related commercial software• Mobile agents vision

Mobile Agent Vision

“...There is a growing danger that agents will be a deception and an empty promise...”

Benn Schneiderman

Mobile Agent Vision

“...agent technologies are most useful when presenting a simpler abstraction of the environment to the user…”

Cybenko and Brewington

Mobile Agent Vision:Justification

• Rapid (re)deployment of applications

• Automated work

• Ability to find and filter information

• Support for push

• Customized views of information

• Support for disconnected operations

• Active network load balancing

Mobile Agent Vision:Student Registration

Assumptions:– Student registers in five courses– When arrives, receives a laptop– Upon first connection to the network, five

course agents arrive at her/his laptop

Mobile Agent VisionCourse Agent

• Each course agent represents a single course

• The user interacts with the course agent through an activity-based user interface

• Standardized interfaces can be customized (through wizards or an agent shell)

Mobile Agent VisionActivity-based Interface

Translators Agent

Contact Instructor

Read Notes

Get Marks

Mobile Agent VisionActivity-based Interface

Translators Agent

Contact Instructor

Do Assignment 1

Read Notes

Get Marks

Mobile Agent VisionAssignment Agent

Assignment 1 on Translators

Read

Do

Submit

Get similar

Mobile Agent VisionMarker Agent

Assignment 1 on Translators: John Doe

View

Execute

Mark

Mobile Agent VisionDiscussion

• Agent bureaucracy provides a functional organization (course related agents, etc.)

• Agents come built-in; can be further customized (explicitly, or implicitly: adaptive agents)

Mobile Agent VisionCollaborative Work

• Two modes of operation:– serial; here the agent moves to the next group

member and brings updates.

Group members may be on- or off-line

Mobile Agent VisionCollaborative Work

• Concurrent: two modes of reintegration of data:

merging versioning

Mobile Agent VisionOn Line Testing

• A test agent can be broadcast to the test group

• Each test agent can control the test, mark it, and then store results somewhere

• The test agent may be “customized” to remember past inputs, etc.

Mobile Agent VisionDiscussion Groups

• Two kinds:– off-line (traditional); for push (Informant)– on-line (chat rooms). You can create an agent to

be your proxy in a chat room

Mobile Agent Vision:Distributed Information Systems

(Work in progress: D. Currie)

• Each IS can be distributed (DIS)

• Each DIS is a directed graph; leaves store information, other nodes classification

• Remote information nodes are kept in cache

• Cache is cleared based on priorities

Mobile Agent Vision:DIS

• Cached items can be marked bad

• Multiple views of the classification can be maintained

• Classification can be built explicitly or implicitly

Mobile Agent Vision:DIS + Agents

• Agents can maintain priorities

• Agents can find and filter information so that it is copied into cache and available off-line

Conclusion

Mobile agents can automate many tasks in mobile computing.

Acknowledgments

Karine Blouin

Yao Chen

Duane Currie

Ali Elbashtiri

Soon Ping Phang

Eng Kian Tian


Top Related