Next-Generation Educational Software Andries van Dam Brown University and the NSF STC for Graphics and Visualization Ed-Media 2002 June 27.

Next-Generation

Educational Software

Andries van Dam

Brown Universityand the

NSF STC for Graphics and Visualization

Ed-Media 2002June 27

Usage Agreement

•These slides may be downloaded and used iff they are correctly attributed and the user sends Andy van Dam ([email protected]) email describing where and how they are to be used.

•If any modifications or corrections or additions are made to these slides, the user agrees to send these changes to Andy van Dam with full permission to use them.

Acknowledgements

•Sponsors

– NSF, Sun, Macromedia, Microsoft, Taco

•Contributors

– Rosemary M. Simpson

– Anne M. Spalter

– Sascha Becker (biology demo)

* Julian Wong, Jim Rusconi, David Kelley, Michael Kowalski, Mark Oribello

Roadmap

•Motivation

•Microworlds

– Exploratories

– Clip Models

•Research agenda

•A modest proposal

•Bulletin: education one of top grand challenges at CRA Grand Challenge workshop

Motivation

•Society mandates– lifelong learning

– just-in-time learning

• Skills needed– adaptability – learning to learn

– critical thinking

– insightful problem solving

• Technology provides– commodity hardware

– federation of devices

– Web-based access, search, and collaboration

Thomas Edison• Motion pictures as educational tool

– “Scholars will soon be instructed through the eye.” - 1913– “I believe that the motion picture

is destined to revolutionize our educational system and that in a few years it will supplant largely, if not entirely, the use of textbooks.” - 1922

Jesse Jesse

Edison’s Legacy

• Bell Labs science films (1950s)

– Frank Capra, director– high school as target audience– superlative and memorable exemplars

• Hemo the Magnificent

• Our Mr. Sun

• Fantastic Voyage

Benjamin Darrow

•Radio as educational tool:

“radio may come as a vibrant and

challenging textbook of the air.”

- 1932

* Founder, first director of the Ohio School of the Air

* “Radio: the Assistant Teacher”, 1932

Alan Kay

... Montessori’s idea was that school should always be an extended kindergarten and it’s the job of people who design the kindergarten to make what happens when kids use it for their own reasons more interesting than the regular world is. I think that's an excellent way of thinking about designing a learning system.”

Brown/MIT Vannevar Bush symposium (October 1995):

http://www.cs.brown.edu/memex/Bush_Symposium_Panels.html

Seymour Papert

•Computer + student choice as educational tool:

“Trying to Predict the Future”, Popular Computing, October 1984

“There won’t be schools in the future…. I think the computer will blow up the school. … but this will happen only in communities of children who have access to computers on a sufficient scale.”

And Other Pioneers• John Anderson

– ACT

• Donald Bitzer

– Plato

• Alfred Bork

– physics

• Alan Borning

– ThingLab

• John Seely Brown

– West and Sophie

• Gregory Crane

– Perseus

• Marc Brown & Robert Sedgewick

– BALSA

• Andrea diSessa

– Boxer

• Doug Engelbart

– NLS

• Hermann Maurer

– Hyperwave

• David Merrill

– TICCIT

• Ted Nelson

– hypertext, Dream Machines

• Bruce Sherwood

– physics

• Patrick Suppes

– logic

• …

Success Stories in Science/Math

Graphing calculator

Geometer’s sketchpad

for Physics teaching

Mathematica

ACT, PAL, PACT cognitive tutors

Vpython

Yet, for each new technology

{euphoria and hypestruggle to produce material for the new mediummature judgmentdisappointment and cynicismwait for the next technology that will be the answer

}

-anon.

Why Now’s The Right Time To Be Excited Again

•Moore’s law commoditizes the necessary platforms- diversity: from graphing calculators to dynabooks

– hypermedia, 3D graphics will be ubiquitous

– decent (multi-modal) UIs possible

•Higher B/W networking, Web distribution– tools multiplying, moving down from professionals

– flexible, extensible roll-your-own courses, JIT learning

– but too much of today’s content is just repurposed, not next-generation that takes advantage of the new capabilities

- educational learning-object consortia, e.g.,NSF’s NSDL and EU’s Ariadne

Genres

Lectures both synchronous & asynchronous

Classical CAIBork, Plato

Traditional multimediaVoyager

Intelligent CAI/tutoringPACT, Biologica

Student programming & scripting

Logo, Squeak, Alice

SimulationsThe Sims, Simxxx

Social simulationsEngines for Education

Multi-player games & worlds SimCountry, MooseCrossing

Our Pedagogical Stance• Constructivist: learning by doing

– Montessori* self-paced, structured environment with self-revealing materials

– Piaget* developmental stages and age-appropriate concepts

– Papert* learning powerful ideas by programming

– Kay* Dynabook -> Squeak

• Spiral approach– Bruner – “…any subject can be taught effectively in some

intellectually honest form to any child at any stage of development.”

• Multiple points of view and goals

• Multiple learning styles (Gardner)

Roadmap

•Motivation

•Microworlds

– Exploratories

– Clip Models

•Research agenda

•A modest proposal

Microworlds: Exploratories

•Simulation-based applets– geometric structure

– behavior

– parameters that are exposed through interfaces* to control and experiment

D

• Made up of re-usable software components

• Web-based delivery of applets embedded in a hypermedia context

• More than 50 in Computer Graphics alone

• Highly interactive

Roadmap

•Motivation

•Microworlds

– Exploratories

– Clip Models

•Research agenda

•A modest proposal

Microworlds: Clip Models

•Simulation-based model

– geometric structure

– behavior

– parameters that are exposed through interfaces

* to control and experiment

– correctly reflect ontology and semantics of domain

•“Clip” suggests ready-made objects for easy embedding in different contexts

– e.g., hypermedia explanation

– also want to combine clip models

Family of Clip Models

•One clip model/concept doesn’t suffice: need a family – levels of detail

* e.g., from anatomical to microscopic views of kidney

– levels of sophistication* e.g., from Bohr atom to quantum mechanical atom

•Additional variations in either model or view/controller– learning styles

* e.g., verbal/textual, visual (1D, 2D, 3D, …), kinesthetic, …

– pedagogical strategies* lecture aid, sandbox, guided exploration, laboratory, …

– device types* e.g., from handhelds to immersive virtual reality

Families of InteroperableClip Models (1/2) •Clip models can be stand-alone but often will be

composed and combined (edu-beans on steroids)– hierarchical composition

* e.g., cardiovascular system is composed of heart, arteries, veins, blood, …

– associations of peers

* e.g., heart, lungs, brain, kidneys, … for respiration

– must import, export suitable parameters

•Difficult to compose models that weren’t co-designed– paucity of collections of reusable learning objects

– meta-data is necessary but not sufficient

Families of InteroperableClip Models (2/2)

•Killer problem: how to provide adaptive focus+context

– allows learner to drill down or explore at will

– interoperability among models of different levels of sophistication and, potentially, simulation,

* e.g., how does senior-high school heart model interoperate with fifth-grade lung and vascular system models?

– plethora of languages and specialized tools

* e.g., Java, Squeak, Flash, Excel, Alice, Maya, …

•Requires design for interoperability

– software engineering plus instructional design plus …

Additional Complexities

•Stand-alone and/or embedded within multiple contexts (e.g., lecture, lab, e-book…)

•Adapt to different and as yet undefined learning environments

– synchronous and asynchronous

– virtual and real classrooms

– on-demand and collaborative learners

– spontaneous study groups

– …

Luxo Jr.

•Pixar award-winning animation

Level of Geometric Detail (and material spec)

Mousing over models brings up menu-based parameter

settings

• Model with sci-viz feature: red shows temperature gradient

• Model with wiring exposed

Bulb color

Bulb wattage

Electricity (ON or OFF)

…

Inputs

Light intensity

Light color

State of (ON or OFF)

…

Outputs

Light Bulb Is Also A Clip Model…

outer geometry simple interior more realistic

structural detail

microscopic view of heated filament

Emits light

Setting Up a Clip Model

Geometry LOD

Structural LOD

Bulb function

Bulb wattage

…

LAMP

High

Med

Low

One-way

Three-way

User configurable

Moving parts

Wiring Just geometry

Show electrical

Show thermal reaction

Show photons

Emits light

Using the Clip Models

• Turns on light and sees different temperatures of light create different spectral breakdowns

•Student wires lamp and chooses filament materials•Adds clip models– a power source (the outlet)

– a prism

Demo: a clipmodel “storyboard”

Project leader: Sascha Becker Brown 1997 A.B. in Computer Science, with honors

1997-2001 software developerConstruct Internet DesignQuokka Sports

Mills College 2002 Post-Baccalaureate Pre-Medical program

Applying to medical school for 2003

Tour through a Digital Human – all faked with a variety of tools (Flash, Director, Maya,…)

Roadmap

•Motivation

•Microworlds

– Exploratories

– Clip Models

•Research agenda

•A modest proposal

What you saw in this demo

•Descent through a structural hierarchy– muscle->cell->myofilament->myosin head-

>mitochondrion

•Multiple modalities– schematic and realistic 2D and 3D

– passive observation and interactivity

•Multiple levels of sophistication– Myosin contraction

– Krebs cycle

•Multiple levels of detail– Krebs cycle

– altitude demo

•Interoperability at different levels of detail

What you didn’t see

•Real models– no simulation

* approximate geometry* faked behavior

– families, e.g., for the heart, faked

•Real interoperability – among simulations at the same level

– among simulations at different levels

•Real navigation

Clip Models - LOD

What would it take to turn this demo into reality? (1/4)

•Instructional design – user point of view– multiple access points

* asynchronous - from search engines, link lists, browsing, …

* synchronous – from coherent curriculum

– different levels of detail present at same time* global contexts plus multiple local focus + context* selective spatial and logical (level-of-detail) zooming

– seamless interaction* bi-directional travel between models* saving state – taking your work away with you* synchronous and asynchronous tele-collaboration* smooth transition to other modes of work

What would it take to turn this demo into reality? (2/4)

•Instructional design – technical point of view (1/3)

– ever-expanding family of clip models for each concept

– interoperability across families for clip models at different levels

– exchange information among the components in unanticipated ways – how to anticipate what sorts of parameters might need to be exposed

* this interoperability problem is what requires more than just standard components

– frameworks for assembly

What would it take to turn this demo into reality? (3/4)

•Instructional design – technical point of view (2/3)

– solving the interoperatbility lies at the intersection of

* simulation science* software engineering* ontology and semantics building

– also involves deep issues of…

* instructional design* user interface design* adaptive hypermedia/agents* …

What would it take to turn this demo into reality? (4/4)

•Instructional design – technical point of view (3/3)

– may be impossibly difficult combinatorial explosion to solve in general

* necessitates domain-specific approaches that weaken generality of concept

– brute force technique

* implement all necessary parameters for every model at deepest level

* almost certainly computationally intractable* requires mixing and matching of stubs and real data for

parameter slots

Grand Challenge-scope project (1/2)

•Digital Human project excellent start* their ontology currently concerned with anatomy, has

to go down to the cellular level

Grand Challenge-scope project (2/2)

•Clip models are embedded in a context

– how do you design persistent learning environments that immerse learners for long periods of time in content?* time on task is most correlated with performance* like learning a language by immersion in the country

– what interfaces encourage interactivity and team learning?

– how do you embed assessment into learning objects?

– how do we measure critical thinking and problem solving?

– how do we build learning environments where students can self-reflect, or talk to experts, or get prolonged mentoring?

– how do we teach conceptual learning?

Roadmap

•Motivation

•Microworlds

– Exploratories

– Clip Models

•Research agenda

•A modest proposal

Learning Federation (1/3)

•Will this massive undertaking happen without external stimulation?

•Neither the government nor industry will do it as a matter of course

– government underspends horribly on education R&D

* DOD does invest in R&D for training

– industry is no better and has been hugely impacted by dot.bomb

– unlike video game improvements, this will not happen without some sort of industry-government planning and funding

Learning Federation (2/3)

•Non-profit, Sematech model – industry-led, industry-government-university collaboration– funds research partnerships

* interdisciplinary teams* old-style ARPA/ONR management

– precompetitive basic and applied research* prototype next-generation learning environments and “courses”* identify and test concepts to be moved into practice

– provide flexible and efficient management* address intellectual property issues* assess ROI * help disseminate result* encourage formation of new research teams and consortia

Learning Federation (3/3)

•Research areas

– learning science (perceptual, cognitive and social science)

– content and instructional design

– evaluation

– assessment

– learning tools and technology

– learning management systems

– monitor innovations in hardware and software

Where is Learning Federation Now?

•Still in formative stage

– initial funding by several companies and NSF for planning and research roadmap effort

– initial research roadmap at NSF in November, 2000

– eight others planned

– fruitful discussion with Larry Grossman re “Digital Promise”/DO IT

“To Infinity and Beyond…”

Resources• Brown University’s Exploratories Project

• Digital Promise/DO IT (Digital Opportunity Investment Trust)

• Digital Human project

• Ariadne Foundation

• CRA (Computing Research Association) Grand Challenge position paper

• “Reflections on Next Generation Software” by Andy van Dam

• “Speculation on a Future Learning Environment” by Randy Hinrichs and Henry Kelly

The End