Software Development Contexts

• product development, where the developers are known in advance but the users are not
• contract development, where the users are known in advance but the developers are not
• in-house development, where both are known in advance

The "waterfall method" of software development, originated in the 1950s, still in use today (vis. the Gateway2000 ordering system, 1996)
1) feasibility stdy, 2) requirements analysis, 3) preliminary design
4) detailed design, 5) coding, 6) integration,
7) installation, and 8) maintenance

Note: these are "stage models" with little flexibility, an emphasis on scheduling, and no emphasis on iteration or prototyping. Why? Because this model was developed for "contract development".
Note also: this approach requires users to "sign off" on a system without benefit of seeing it or using it.

• developers and users are separated (by space, time, and organization)
• systems were not interactive

Why not stress interation and prototyping in design?
"It will cost too much and take too long"

The cost-benefit and business case analysis of usability testing and usability engineering is leading towards viewing usability flaws as software bugs (that NEED to be fixed).

Interactive Systems: Bridging the Gap Between Developers and Users, by Jonathan Grudin (RiCHI pp. 293-303)

The "waterfall" method of software development continues to be used in contract development contexts, where it is natural

• because projects are defined before developers are identified
• it has been successful in non-interactive systems
• but, with interactive systems it creates a "wall" between users and developers (because they follow spec, not user, guidelines)

Product development focuses on functionality over usability

• because a product can be successful by meeting "the usability needs of only a fraction of the potential users"
• the human-computer interaction communicty has had the largest effect on this group

In-house development is the smallest and yet most influential in terms of recent trends to usability engineering

Each of these paradigms have their own "opportunities" and "obstacles" to which Grudin adds "mediators":

• Mediators in contract development are outside consultants paid to represent users in the development cycle
• Mediators in product development are the sales and marketing departments, and outside consultants paid to represent users
• Mediators in in-house development are the end-users (if they're involved in the process)

The message: the working context of development is another factor influencing user interface design and usability engineering.

Conducting and Analyzing a Contextual Review (Excerpt)

"Contextual inquiry does not provide a set of steps to follow for collecting and interpreting user information. Rather, it provides a set of concepts that guide the design and implementation of information collection and analysis sessions."

Contextual Inquiry "challenges"

• to support design across different contexts and cultures
• to gather user information under time and resource constraints
• to gain useful information within a working context

The REAL Problem: blind faith in users who:

• do not know what they want and/or do not remember what their problems are
• have experience with their work, but no experience with designing systems
• express opinions, when pressed, that are divorced from their experiences (because they are out of context)
• There is a significant difference between summary information and ongoing experience

Advantages of contextual inquiry:

• supports production of general purpose systems
• provides a way to work with users for a SHORT PERIOD OF TIME
• fosters team work and shared vision
• fosters participatory design

How it works:

• develop models of current work practice using interviews with users
• then uses these models to design a system model

How it REALLY works

• Designers watch people doing work
• Designers propose potential improvements with "paper prototyping"
• Interviews are "focus" oriented, not question-driven
• Interviews are unstructured and depend on good "people skills" like listening, noticing non-verbal cues (and sending thank you notes!)
• Pg. 247-248 give a DETAILED, step by step, account of how to conduct a contextual interview: from introducing yourself, to thanking them for their time

Interpreting/Analyzing interview results:

• Transcribe the interviews
• Enter into a group dialog with the transcription
• Develop an Affinity Diagram

Affinity Diagrams (aka Cognitive Maps)

• An attempt to organize and categorize unstructured knowledge
• First step in what is called Knowledge Acquisition in AI
• Danger of creating informal maps which don't answer the requirements questions

Controversies about Computerization and the Organization of White Collar Work by Rob Kling

• Assumption: technologies have fundamentally changed the nature of work
• Assumption: working life is important to social life

History

• Both blue- and white-collar work is becoming computerized
• Office work has become much less "mechanical"
• Systems have become increasingly powerful and complex
• Data processing is no longer a specialized department

Social Issues

• "Progress" is open to interpretation
• Utopian viewpoints describe modern computer work as creative and interesting
• Anti-Utopian viewpoints describe a closed society of technocrats changing society for the worse

Social Realism (a data driven account)

• Gender Issues: 80% of the clerical workers are women, less than 30% of the programmers are women (trend: in 1966 0.4% engineers, in 1987, 14%)
• Control Issues: Electronic workplaces are increasingly "monitored"
• Control Issues: Information workers have, temporarily at least, increased flexibility in their work and working conditions
• The Skilling Debates:
  • deskilling: reducing the skills required for jobs, in order to hire less expensive workers
  • automation causes job-shifting
  • powerful software (CASE tools) continue to change the CS profession
  • it is unlikely that new technologies alone will improve work
• Culture determines design: which explains why inter-cultural methods (like Contextual Review, by Holtzblatt and Jones) are increasingly important