NPTEL Course on

Human Computer Interaction

- An Introduction
Dr. Pradeep Yammiyavar Dr. Samit Bhattacharya
Professor, Assistant Professor,
Dept. of Design, Dept. of Computer Science
IIT Guwahati, and Engineering,
Assam, India IIT Guwahati, Assam, India

Indian Institute of Technology Guwahati

 Module 2:
Interactive System Design

Lecture 2:
HCI and Software Engineering

Dr. Samit Bhattacharya

 Objective

• In the previous lecture, we have learned about

the idea of “usability”

– While designing an interactive system, what we

should do to take care of usability?

• In this lecture, we shall learn about the

answer to the above question
 Objective

• In particular, we shall learn about the

following
– The difference between a software design and an
interactive system design

– User-centered and participatory design

– The interactive system design life cycle

 The Central Idea

• Suppose you are designing a database

management system (DBMS): what are your
design objectives
– Efficient storage of large databases (storage)

– Efficient way to retrieve results of a query from

database (retrieval)

– Allowing the user to access the database

(interaction)
 The Central Idea

• Note that this is a scenario where the user

interacts with the system (database)

• However, the user is a “computer expert”,

who has “technical knowledge” about the
system
– Through some query language, the user can
access, manipulate and update the database
 The Central Idea

• Now consider a tourist information system

• In the background, it is nothing but a database

of various tourist-related information

• However, its users may or may not be

“computer experts”
– They do not care about what goes on inside
– They just want to “get” the information “easily”
 The Central Idea

• The term “easily” is very significant

– It means, we need to have an interface and
interaction mechanism that do not require any
specialized knowledge

• That is, we need a “usable” system

• Design goal of an interactive system:

increase usability
 What Happens in Software
Engineering

• The waterfall model: the simplest and typical

way to visualize software design

• Design process composed of a series of sub-

stages
– Each sub-stage follows the previous stage and
precedes the next stage (looks like a waterfall)
 The Waterfall Model

Requirements
specification

Architectural
design

Detailed
design

Coding and
unit testing

Integration
and testing

Operation and
maintenance
 What Happens in Software
Engineering

• Note the uni-directional flow (that’s how real

waterfalls work anyway!!)

• In other words,
– Each stage depends on the previous stages but not
vice-versa
 Interactive System Design

• The uni-directional flow is not appropriate for

interactive system design

• In other words,
– Each stage depends on the previous stages. It may
also depend on the next stages (feedback).

• It is no longer the (linear) waterfall model

Interactive System Design

Requirements
specification

Architectural
design

Detailed
design

Coding and
unit testing

Integration
and testing

Operation and
maintenance
 Why This Difference

• We are trying to design “for” the user

– Not for a programmer’s convenience or expert’s
use

• What should we consider

– Abilities and needs of the users
– Their usage context
– Their work setting

• In other words, we have to “know the user”

 Need: Know The User

• A never ending process because there is so

much to know and because the users keep
changing

• An interactive system designer should

consider the human factors that characterize
users
 The Human Factors

• Perception: our ability to perceive our

surroundings
– Can be visual, auditory or haptic (touch)

• Cognition: the way we process the perceived

information in our “mind” and take decisions
• Motor action: this is the mechanism through
which we interact with the surrounding
– Example: hand movement, eyeball movement,
speech
 Need: Know The User

• These factors (user characteristics) vary with

– Age, gender, physical and cognitive abilities,
personality
– Education, cultural or ethnic background
– Training, motivation, goals

• An interactive system designer should recognize

this diversity
 Need: Recognize Diversity

• Systems used by several communities of users

– No single design can satisfy all users and situations

• Designer faces real challenge to cater to the

need of each community
– Designers must characterize users and situations as
precisely and completely as possible
A Generic User Characterization

• Novice or first time users

– Know nothing about the task or interface concepts

– Often anxious about the computer and its

functionality

• Knowledgeable or intermediate users

– Have stable task concepts and broad interface
concepts
A Generic User Characterization

• Expert users
– Thoroughly familiar with the task and interface
concepts

– Want to get their job done quickly

 So, Why The Difference?

• Designer must know the user

– This knowledge can not be captured at once

• Design involves acquiring new knowledge and

using it to refine design in continuous cycle
(till some “acceptable” design is found)
– The reason for so many “feedbacks” in the waterfall
model
User Centered Design (UCD)

• The design process, where designer collects

feedback about the design from users and use
this to refine design, is known as “user
centered design” or UCD

• UCD is based on understanding the domain of

work or play in which people are engaged and
in which they interact with computers
User Centered Design (UCD)

• Assumptions
– Result of a good design is a satisfied user
– Process of design is a collaboration between
designers and user.
– Design evolves and adapts to users’ changing
concerns, and the process produces a specification
as an important byproduct
– The user and designer are in constant
communication during the entire process
 UCD Drawbacks

• In UCD, user involvement is “passive”

– The designer elicits feedback from user (through
interviews, informal discussions etc)

– Prepares specification on the basis of user response

– Take feedback on the design and makes refinements

 UCD Drawbacks

• Problems with “passive” involvement of user

– User intuition about a new interface may not be
correct (feedback not reliable)

– The interview process itself may not be formulated

properly (designer asks wrong questions)

– It is not possible for the designer to identify all

possible issues to take feedback from users, as the
designer’s knowledge about the user may not be
complete
 Participatory Design

• Solution: make (representative) users a part of

the design team

• Such a design process, where end users are part

of the design team, is known as “participatory
design”
Participatory Design: Key Points

• Users are first-class members of the design

team
– As opposed to their passive involvement in UCD

• Users are considered subject experts

– Know all about their work context

• Iterative design process

– All design stages are subject to revision
Interactive System Design Life Cycle
(ISLC)

• Key stages
– Know the user, propose design, evaluate design by
users, refine design

• Iterative design process

– The above stages are iterated till an acceptable
(determined from user feedback) solution is
obtained
 ISLC:
A Consolidated Diagram

Identify
need
Analyze
data
Propose
design
Implement
and deploy
Evaluate Develop
Prototype
 Life Cycle Stage:
Identify Need
• What is wanted – identify users and their need

• Designers make use of one or more methods

to identify the requirements
 Life Cycle Stage:
Identify Need
• Such methods include
– Interview (structured, semi-structured,
unstructured)
– Contextual inquiry
– Cultural probes
– Ethnography
– User models
 Life Cycle Stage:
Analyze Data
• Analysis of the data collected

• Two types of analysis are performed

– Scenario analysis: analyze data collected

from the user on one or more usage
scenario of the system
 Life Cycle Stage:
Analyze Data
• Analysis of the data collected

• Two types of analysis are performed

– Task analysis: analyze tasks required to be

carried out by the user to operate the system
• System level task analysis: analysis of external
tasks required to operate the system
 Life Cycle Stage:
Analyze Data
• Analysis of the data collected

• Two types of analysis are performed

– Task analysis: analyze tasks required to be

carried out by the user to operate the system
• Cognitive task analysis: analysis of tasks
performed in the mind of the user
 Life Cycle Stage:
Propose Design
• Design proposal arrived at from the analysis
of collected data

– Guidelines and principles help in the

development of initial design

– Several sets of guidelines (both general and

specific) are there to cater to specific
interface design context
 Life Cycle Stage:
Develop Prototype
• Implement a prototype of the design for
collecting user feedback
• A spectrum of techniques is used in
developing prototypes
– Paper prototype (one extreme)
– Complete software (other extreme)
– Lots in between …
 Life Cycle Stage:
Evaluate Design
• Evaluation of the design by users

• In the initial design phase, evaluation is done

on prototypes
– Cost effective and easier to perform

– Suitable for iterative design process where the

evaluation is performed many times
 Life Cycle Stage:
Evaluate Design
• Evaluation of the design by users

• The full system is typically evaluated at the

end
– Full system evaluation is costly in terms of money,
manpower, time and effort

– Hence, typically done once or a limited number of

times
 Life Cycle Stage:
Evaluate Design
• Several evaluation methods are available

– Checklist/guideline based evaluation

• Heuristic evaluation, cognitive walkthrough

– Model-based evaluation: employs models (of the

system or user or hybrid) for evaluation
• Hybrid models are essentially models that
combines the features of both the system and
the user
 Life Cycle Stage:
Evaluate Design
• Several evaluation methods are available

– Empirical evaluation – evaluate with real

users
• Involve implementation of the system with full
functionalities
 Note

• In subsequent lectures, these stages and

methods of the ISLC will be discussed in
details