Supporting Life Scientists via End User Programming

1
Supporting Life Scientists via End User
Programming

• Luke Church
• Computer Laboratory, University of Cambridge
• Microsoft eScience - Dec 08

With thanks to co-authors
2
Agenda

• End User Programming (EUP)
• What is it?
• How is it being used?
• Supporting Scientists
• 1 Usability
• Applying EUP to improve our tools
• State of the art Cognitive Dimensions of
  Notations
• 2 Executable Biology Life as Computation
• Successes
• Challenge natural abstractions

3
End User Programming?
Mainstream Usability
Direct Manipulation
Programming
Actions on lots of objects Actions when Im not
here Communicating intent Abstraction
Actions on objects
Current UI paradigm
Professional Programmers
End Users
End User Programming
Power, Expressiveness, Communication
4
End Users Can Program
Given the right problem and tools
A few million users...
5
End Users Can Program
Given the right problem and tools
6
End Users Can Program
Given the right problem and tools
(From http//blogoscoped.com/archive/2007-02-10.ht
ml)
7
End Users Can Program
Given the right problem and tools
8
Why give EUP a name?

• Programming is a different cognitive task
• Direct Manipulation usability methods dont help
  much

Reasonable from a direct manipulation
perspective
But has Hidden Dependencies poor
Provisionality poor Progressive
Evaluation etc.
9
Why give EUP a name?

• Programming is a different cognitive task
• EUP analysis tools acknowledge this
• By understanding that something is programming
• Improved design discussions
• Text or graphics?
• Impact of type systems?
• More appropriate usability techniques
• Cognitive Dimensions of Notations
• Attention Investment Framework
• User Experience of Intentional Information

10
EUP for life Scientists1 Usability
11
Modelling tools

• A particular type of programming language
• We used analytical EUP techniques to improve
  usability
• Small user study showed a good provisional result
• Prototype now in refinement

12
Cognitive Dimensions

• State of the art usability technique
• Consider an environment, along a number of
  dimensions
• How defuse is the information?
• How viscous is it?
• How many abstractions does the user have to
  learn?
• Does the user have to prematurely commit to a
  structure?
• Highlight areas of likely problems, suggest
  design manoeuvres

13
Cognitive Dimensions
14
Successes

• Identify and fix deep usability problems
• Academically verified and supported
• 50 extension papers, widely cited within
  community
• Theoretical testing, practical adoption
• Workshops, tutorials, handbook in progress
• Increasing adoption
• Scientific tools
• API design at Microsofts
• Security usability
• Spreadsheet UI design
• ...

15
EUP for life Scientists2 Executable Biology
16
Executable Biology

• Executable models of living processes

Compare predictions to data
Suggest new experiments
Adjust model
17
Executable Biology

• Successes
• Experimentally verified refinements in regulatory
  models of various organisms
• Proposed new concurrency models to reflect
  biological processes
• Successful application of formal verification
  techniques to laboratory data
• Challenges
• Challenges laboratory practice around timed
  measurements in experiments
• Develop computational models that are more
  suitable for biological modelling

18
Executable Biology

• New Abstractions?
• Need for abstraction processes that work with
  biology
• Translucent, socially supported, computationally
  tractable
• Abstraction problem in computers, let alone
  biology
• Possibility for great mutual benefit
• What does it mean to understand a program?
• Pragmatic computationalism, it is useful to
  consider life ? computational model
• What does it mean to understand a program
  independent of a goal?

19
On Abstraction...

• Abstraction carries Intention
• Classifications/Abstractions are a shaping
  perspective
• E.g. International Classification of Disease
• Abstracting biological systems gt shaping what we
  can know
• What questions are interesting?
• Traditionally decided by the scientific community
• Accelerating time to discovery?
• We are deciding what questions are answerable
• Adoption requires supporting social processes
• Responsibility to make answers useful

20
Summary

• Usability
• Usability is crucial to adoption and efficiency
• Standard techniques necessary but not sufficient
• End User Programming techniques help
• Executable Biology
• Pragmatic computationalism
• Understand living systems like programs, positive
  initial results
• Deeper challenges
• What does biology mean by understand?
• Provide new abstraction processes, with social
  support
• Accelerating discovery by providing new ways of
  thinking about biological systems