eu_20051006_euses - College of Engineering | Oregon State

Carving up the Space of
End User Programming
EUSES, Lincoln, NE, Oct ‘05
Agenda
• Overview of our conceptual “space”
(The projects below are aimed at refining this space.)
• Past projects:
– Re-analysis of 55M data
– Survey of Information Week readers
• Future projects:
– Interviews of Katrina-related web & db developers
– Contextual inquiry of specific populations
Purpose of our “Conceptual Space”
• Our goal is to understand the population of
end users (EUs) who program.
– What are EUs’ strengths and weaknesses?
– What sorts of programming are they doing?
– How many EUs are doing each type of programming?
– Where can we invest our time to achieve significant benefits?
• Answering these goes hand-in-hand with
mapping out our “conceptual space.”
Where are the strengths of EUs?
Means of
Programming
Activity Type
Task Structure
55M / 90M estimate & task structure
• Updated 55M estimate
– Old: 55M EU programmers in 2005
– New: 90M EU in 2012
– New: incl. 55M spreadsheet and/or db users
• Received insight into most common tasks
– Most common occupations for EUs:
Manager, teacher, secretary, accountant
Results reported in C. Scaffidi, M. Shaw, and B. Myers. Estimating the Numbers of End Users and End
User Programmers. Proceedings of VL/HCC, 2005.
More focused “task structure” dimension
Means of
Programming
Activity Type
Task Structure
- Accountant
- Manager
- Secretary
- Teacher
- (Others)
Survey of feature usage
• 2005 survey of Information Week readers
– Ask about usage of application features
– Focus on abstraction-related features
(E.g.: JavaScript, web server scripting, databases, macros, and
spreadsheet features)
• Propensities to use features fell cleanly into
three clusters
– Macros, Linked Structures, Imperative Code
Results to be reported in C. Scaffidi, A. Ko, B. Myers, M. Shaw. Identifying Types of End Users: Hints
from an Informal Survey, Technical Report CMU-ISRI-05-110/CMU-HCII-05-101, Institute for
Software Research International, Carnegie Mellon University, Pittsburgh, PA, 2005.
More focused “means” dimension
Means of
Programming
- Macros
- Linked Structures
- Imperative Code
- (Others)
Activity Type
Task Structure
- Accountant
- Manager
- Secretary
- Teacher
- (Others)
Moving along to future projects…
• Two past projects refined our “space”
– Re-analysis of government data helped refine “task” dimension
– Information Week survey helped refine “means” dimension
• What about the “activity” dimension?
– Katrina-related “person locator” study
– Contextual inquiry of three populations
Study of Katrina-db creators
• Fall 2005 telephone interviews
• How do different EUs handle one need?
– Need: “person locator” site
– Solution: wide and varied, depending on EU
(Some are even syntheses of existing web databases.)
–
–
–
–
How did they decide what to build?
Why did they decide to build in the first place?
What types of activities were difficult?
How did they overcome these difficulties?
Cross-cut of web and db “means”
Means of
Programming
- Macros
- Linked Structures
- Imperative Code
- (Others)
Activity Type
(e.g.: knowledge,
comprehension,
application, analysis,
synthesis, evaluation)
Task Structure:
- Accountant
- Manager
- Secretary
- Teacher
- (Others)
Study of data interoperability problems
• Fall 2005 contextual inquiry
• How do different EUs cope with problems?
– Focus: data interoperability between apps
– Population:
• Administrative assistants / secretaries
• Managers (emphasis on marketing managers)
• Graphic designers (intended as a half-step toward
professional programmers)
– Hopefully we will gain insight into how Linked Structure features
assist or confound EUs.
Study inspired by article “Science fiction?” in The Economist, Sep 2005.
Cross-cut of linked structure “means”
Means of
Programming
- Macros
- Linked Structures
- Imperative Code
- (Others)
Activity Type
(e.g.: knowledge,
comprehension,
application, analysis,
synthesis, evaluation)
Task Structure:
- Accountant
- Manager
- Secretary
- Teacher
- (Others)
Summary
• Past Work
– Extending the EU count estimate
– Scoping out most common EU occupations (“task” dimension)
– Exploring propensities to use abstractions (“means” dimension)
• Future Work (“activity” dimension)
– Seeing how various EUs respond to one need
– Scoping out data interoperability problems
Thank You
• To the EUSES community for your
interest and feedback
• To NSF, Sloan, and NASA for funding
References
55M/90M estimates: C. Scaffidi, M. Shaw, and B. Myers. Estimating the Numbers of End
Users and End User Programmers. Proceedings of VL/HCC, 2005.
Feature clustering: C. Scaffidi, A. Ko, B. Myers, M. Shaw. Identifying Types of End Users:
Hints from an Informal Survey, Technical Report CMU-ISRI-05-110/CMU-HCII-05101, Institute for Software Research International, Carnegie Mellon University,
Pittsburgh, PA, 2005.
Inspiration for interoperability study: “Science fiction?” in The Economist, Sep 2005.
Bloom’s taxonomy: B. Bloom, B. Mesia, and D. Krathwohl. Taxonomy of Educational
Objectives. David McKay Publishers, New York, NY, 1964.
Green and Blackwell’s activity type categories: A. Blackwell and T. Green. Cognitive
Dimensions of Notations Tutorial at VL/HCC, 2005.