If we build it will they come The challenges of cyberinfrastructure development

1
If we build it will they come? The challenges of
cyberinfrastructure development

• Thomas A. FinholtSchool of InformationUniversity
  of Michigan

2
Outline

• The field of dreams
• Recommendations of the NSF panel
• Challenges
• Individual
• Group
• Cultural
• Prospects

3
The field of dreams
I was sitting on the verandah of my farm house in
eastern Iowa when a voice clearly said to me, If
you build it, he will come. Ray Kinsella in
Shoeless Joe, by W.P. Kinsella
Image source http//www.fieldofdreamsmoviesite.co
m/
4
The Atkins report A brief summary

• Nothing tends so much to the advancement of
  knowledge as the application of a new instrument.
  
• Sir Humphrey Davy

Quotation source Thomas Hager, Force of Nature,
Simon and Schuster, New York, 1995, p 86Image
source Sir Thomas Lawrence, circa 1821, National
Portrait Gallery, London
5
Atkins panel recommendations

• Data as the instrument
• Network as the instrument
• Simulation as the instrument

6
Data as the instrument

• by-products as products

7
Examples

• Past
• public health reporting
• Present
• virtual observatory
• Future?
• car versus deer

8
pump
Source http//www.ph.ucla.edu/epi/snow.html
9
Network as the instrument

• sensors, everywhere, joined

10
Examples

• Past
• Bell system
• Present
• GPS and TEC plots
• Future?
• computational and data grids

11
Global GPS Network (November 1996) Coverage at
Ionospheric Heights
10 degree elevation mask. Intersection height of
400 km.
Source http//iono.jpl.nasa.gov/sitemap.html
12
Source http//iono.jpl.nasa.gov/latest_rti_global
.html
13
Simulation as the instrument

• seeing beyond the field-of-view

14
Examples

• Past
• physical models
• Present
• theory/data closure
• Future?
• multi-scale

15
UARC Simulation and observational data
16
SPARC Simulation and observational data
Source http//sparc-1.si.umich.edu/sparc/central/
page/TomsTINGvsObserved
17

• Synchronous communication
• Asynchronous communication

Researchers

• Synchronized data
• Synchronized data and images
• Data discovery

• Teleoperation
• Teleobservation

Data
Facilities

• Automatic archiving
• Simulation codes
• Hybrid experiments

18
Bhuj, India. One of the towers of this apartment
complex totally collapsed,and the central
stairway leaned on another building of the
complex. Photo courtesy of Dr. J.P. Bardet,
University of Southern Californiahttp//geoinfo.u
sc.edu/gees/RecentEQ/India_Gujarat/Report/Damage/B
huj/Bardet_Feb18.html
19
Shake table Nevada, Reno
20
Reaction wall Minnesota
21
(No Transcript)
22
NEESgrid interface
23
NEESgrid interface
24
NEESgrid interface
25
Aurora borealis
26
Incoherent scatter radar Sondrestrom
Photo courtesy of Craig Heinselman, Stanford
University and SRI Internationalhttp//www-star.s
tanford.edu/craig/facility.htm
27
SPARC interface
28
Instrument scope, 1993
29
Instrument scope, 2001
30
Individual challenges

• Costs incurred by developers are borne by users
  Grudin
• UARC/SPARC
• Burden of NeXT environment
• Worm Community System
• Burden of unix environment
• NEES
• Burden of globus and OGSA

31
UARC interface
32
Total sessions per year
33
Change in system use
34
UARC 6.0 interface
35
Group challenges

• Competency traps March
• UARC/SPARC
• Campaign orientation
• Unequal distribution of participation
• NEES
• Preference for Windows
• Pritzker
• Preference for face-to-face

36
Persistence of the campaign orientation1999-2000
Radar campaign
LTCS campaign
Geophysical world days
Workshop
37
Skewed campaign participation 1993Observed at
Sondrestrom
80 of the users generated 40 of total talk
38
Skewed campaign participation 1993 Observed in
UARC
80 of the users generated 33 of total chat
39
Skewed campaign participation 1999Observed in
SPARC
80 of the users generated 30 of total chat
40
Cultural challenges

• Geography of thought Nisbett
• NEES
• earthquake engineers vs. grid specialists
• Lucent
• west vs. east
• GLRCFAR
• clinical vs. primate

41
Earthquake engineers in Hofstedes scheme

• Power distance
• Hierarchical
• Bias toward seniority
• Individualist
• My lab is my empire
• Solo PI model
• Masculine
• Adversarial
• Competitive
• Uncertainty avoidance
• Highly skeptical of new technologies
• Extremely risk adverse

42
Grid specialists in Hofstedes scheme

• Power distance
• Egalitarian
• Bias toward talent
• Collectivist
• Use the Internet to create worldwide communities
• Project model
• Masculine
• Adversarial
• Competitive
• Uncertainty avoidance
• Extremely open to new technologies
• Extremely risk seeking

43

44
Agreeing on terms
45
How earthquake engineers think
Customer Need
Customer Requirements
Structure Design
Structure Construction
Structure Acceptance
Structure Operations
Customer Needs Assessment
46
How grid specialists think
47
Prospects

• Attempts to apply new technology are often framed
  in terms of familiar technology
• First efforts are often awkward hybrids
• It is hard to know where the seeds of greatness
  might lie...

Charles Kings horseless carriage (1896)
Detroit, Michigan
Source American Automobile Manufacturers
Association, http//www.automuseum.com/carhistory.
html
48
The culture of simulation -- Turkle

• Concrete vs. abstract
• Exploratory vs. rule-based
• Improvisational vs. planned

49
The cottage cheese story

• A diet conscious cook is allowed a half cup of
  cottage cheese per day
• Recipe calls for two-thirds of daily allotment
• Cook forms circle of cottage cheese, divides into
  thirds -- scrapes two segments into mixing bowl

50
Products of the simulation design aesthetic

• What makes a design good?
• Mutability
• Who does the designing?
• just plain folks
• What is a signature design achievement?
• the SARS global virtual laboratory
• hosted on a secure Web site

Good morning, good day, good evening
Image source Enserink, M., Vogel, G. (2003).
Deferring competition, global network closes in
on SARS. Science, 300, 224-25.
51
NEJM editorial April 2, 2003
Use of the Internet has sped information
exchange and helped overcome the problems
presented by asynchrony in the activities of
investigators in many time zones. Scientists at
the international collaborating laboratories are
exchanging laboratory results and images on a
secure Web site. Coordination of the
international response strategy has been fostered
by regular videoconferences with senior leaders
in the operations center at the WHO, the DHHS,
and the CDC. Satellite broadcasts, Webcasts, and
videoconferencing are supporting the
dissemination of emerging information to the
entire global health community.
Source http//content.nejm.org/cgi/reprint/NEJMe0
30067v1.pdf
52
How to tinker
Source http//www.tam.cornell.edu/ruina/hplab/
53
Tinkerers as change agents

• They make sense of the world in light of
  experience
• They need to play with applications to appreciate
  their function
• True requirements may emerge only after false
  starts

54
Tinkering skills

• Empathy -- can you see things through the users
  eyes?
• Flexibility -- can you experiment?
• Plagiarism -- can you find and assimilate
  successful innovations from other systems and
  services?

55
Human-centered tinkering

• Define requirements in terms of observed models
• Test hypotheses in actual communities
• Use feedback to improve systems and services

56
Observe
Conceptualize Observe models
57
Observe, Build
Conceptualize Observe models
Build Intervene
58
Observe, Build, Test
Conceptualize Observe models
Build Intervene
Trials Deploy, use, evaluate
59
Observe, Build, Test, Modify
Conceptualize Observe models
Build Intervene
Trials Deploy, use, evaluate
Modify extend design, evolution
60
Wired VS reality
More
raw performance of technology
hype
Performance
reality gap
real performance
Less
Time
61
Building it so they will come

• Give users objects to think with (scenarios,
  mock-ups, prototypes)
• Be patientlet users convince themselves
• Know where youve been (collect baseline data)
  and whats changed (collect data as you go along)

62
(No Transcript)
63
Building it so they will come

• Create a new interdisciplinary workforce
• p. 26 in the Atkins report 2003
• p. 13 in the Arzberger and Finholt report 2002
• p. 58 in the ETAN report Transforming European
  Science through ICT 1999

Sources Atkins report --http//www.communitytechn
ology.org/nsf_ci_report/ Arzberger and Finholt
report -- http//nbcr.sdsc.edu/Collaboratories/Col
laboratoryFinal2.doc ETAN report --
ftp//ftp.cordis.lu/pub/etan/docs/ict-report.pdf
64
Building it so they will come

• Balance contributions (pp. 50-51 in the Atkins
  report)
• Weight domain science too heavily?
• Overemphasize procurement of existing
  technologies
• Computer scientists become viewed as merely
  consultants and implementers
• Weight computer science too heavily?
• End user needs insufficiently addressed
• Emphasis on novelty at the expense of usability
  and stability

Source Atkins report --http//www.communitytechno
logy.org/nsf_ci_report/
65
Innovation vs. extrapolation
Innovation

• computational grids

• e-science

Social Technological Forces

• collaboratories

• distance learning

• community networks

• electronic commerce

• digital libraries

• electronic journals

• video conferencing

Extrapolation
66
Building it so they will come

• Remove institutional barriers to collaboration
• David and Spence draft report
• p. 13 Arzberger and Finholt report
• OECD report on access to publicly funded research
  data
• NIH statement on data sharing
• Pritzker covenant

Sources Arzberger and Finholt report --
http//nbcr.sdsc.edu/Collaboratories/Collaboratory
Final2.doc OECD report -- http//dataaccess.ucsd
.edu/FINAL_Interim_Report_20_Oct2002.doc NIH
draft statement -- http//grants2.nih.gov/grants/p
olicy/data_sharing/
67
Relevant URLs

• crew.umich.edu
• si.umich.edu
• neesgrid.org
• nees.org
• www.scienceofcollaboratories.org

Shoeless Joe Jackson
Image source http//www.chicagohs.org/history/bla
cksox/joe.html