Title |
Review of Masters Degree in Human-Computer Interaction at SUNY Oswego
|
Author |
Gary Perlman (perlman@oclc.org)
|
Modified: |
|
Abstract |
In early December 2000, I visited SUNY Oswego
to review their plans for a Masters program in HCI
(Human-Computer Interaction).
The planning for the program is at an early stage,
which is an ideal time to bring in outside expertise.
This document is a summary of my impressions from the visit,
and my recommendations for further planning.
My general strategy was to synthesize my notes into a framework
that followed the flow of students in, through, and out of the program.
For a general audience, I begin with a definition of the field of HCI.
I conclude that an HCI Masters program can do well at SUNY Oswego,
but that finding students and faculty will be a challenge.
|
Biography |
Gary Perlman, Ph.D. is a consulting research scientist at the OCLC Online Computer Library Center.
He designs, builds, and evaluates user interfaces for bibliographic and full text retrieval,
notably for the OCLC FirstSearch service.
Before OCLC, he had research and/or teaching positions at Bell Labs,
Wang Institute, Software Engineering Institute, MIT, and Ohio State University.
Dr. Perlman has a Ph.D. from the University of California, San Diego,
and a B.A. from the University of Rochester, both in psychology.
Having been a professor for 12 years,
he has had more than a passing interest in HCI education:
|
Contents: |
|
According to the ACM SIGCHI Curriculum Development Group's 1992 report,
Human-computer interaction is a discipline concerned with the design, evaluation and implementation of interactive computing systems for human use and with the study of major phenomena surrounding them.
While the
design, evaluation and implementation of interactive computing systems for human use
can be viewed as the
engineering practice
of the field,
the
study of major phenomena surrounding them
can be viewed as the
scientific investigation leading to theory.
While the dichotomy between engineering/practice and science/theory is artificial,
it is a natural one to highlight when discussing a Masters program of study
compared to a Doctoral program, especially in a field for which graduates
are in high demand in industry.
An assumption throughout this review is that SUNY Oswego is going to be more
focused on preparing students to
contribute to the engineering of more usable software systems
than more long-term advancement of scientific theory.
The terms "HCI", "HCI practice", "usability", "user interface development" will be used
somewhat interchangeably, with the qualification that good engineering is
based on good science, and often helps advance science.
Software is usable to the extent that is makes it possible for people
to accomplish useful tasks within human limitations and other constraints.
Usable software begins with the identification of the tasks people need to accomplish,
and ends with the evaluation of the degree to which people can actually accomplish those tasks.
Thus HCI practice is inherently human-centered,
from the analysis of requirements, through design, to system evaluation.
Many skills from many disciplines are needed to produce usable software.
Not the least of which is software engineering,
because without that discipline's goals of validity, reliability, performance, cost-effectiveness, etc.,
usability becomes a secondary concern.
But a system that does not meet user needs or task demands,
as might be best identified by psychologists, anthropologists, marketers, industrial designers, etc.,
or if it is not conveyed through effective graphic design,
a well-engineering system may have no use or no users.
HCI is a multidisciplinary field, as evidenced by the range of fields
listed on the web site of its leading professional organization, ACM SIGCHI:
computer science,
psychology,
cognitive science,
human factors,
industrial design,
graphic design,
anthropology,
sociology,
management science,
software engineering,
...
Given the multidisciplinary nature of HCI,
it is important to emphasize that while people from
many disciplines contribute to system development,
and people may contribute more to areas related to their discipline,
none can dominate the contributions in an area.
In HCI, we make decisions based on the
data from the user
(not just user opinions, and not contributor opinions).
Oswego is not is the best location for starting or sustaining a Masters program in HCI.
An HCI program should attract students from industry,
and a major metropolitan area would be better.
Given its proximity,
it will be important to leverage
Syracuse and its high technology sector,
and offering some courses in Syracuse may be considered.
Cities as near as Rochester would be too far for students to commute
more than once or twice a week, making part time involvement in the
program difficult.
Distance education may be acceptable for some courses in the program,
but not for group projects.
A key to promoting the program may be obtaining industrial cooperation,
perhaps in the form of an industrial advisory board.
Through industrial advisors, SUNY Oswego might obtain:
- suggestions about what should be included in the program
- sources of funding
- sources of students
- adjunct faculty
In return, these industrial champions of HCI would have
greatly enhanced input into the program,
influencing it in directions that increase benefits to their companies.
Most students at SUNY Oswego are full time,
but an HCI Masters program should ideally lend itself to part-time
and non-degree (continuing education) enrollment.
It may turn out that
most students will come from industry,
either at the request of their employer,
or based on their own initiative, motivated by dissatisfaction
with the usability of software,
but possibly based on their excitement of increasing usability, accessibility, etc.
Some students will come from the upper level undergraduate ranks,
for which a 3+2 or 4+1 program has been discussed.
Most Masters students will probably be software developers
(with degrees in computer science, engineering, etc.),.
The program may be attractive to
technical writers, marketers, and graphic designers,
and students with psychology or human factors degrees.
Attracting students with varied backgrounds may be critical to the success
of the program because of the multidisciplinary nature of HCI.
Having a team of programmers play multidisciplinary roles is nothing
like working with a real graphic designer or human factors specialist.
Drawing students with diverse backgrounds implies that
the graduates of the program will be diverse:
- programmers will become user interface programmers
- psychologists will become usability specialists
- graphic artists will become visual interaction designers
and so on, each prepared to work with the other in a team,
and each having some skills outside their primary area.
What Background is Needed for Students?
The ACM SIGCHI Curriculum development group proposed the following pre-requisites for its
proposed graduate-level courses in HCI:
All students taking the proposed courses should:
- be advanced undergraduates or beginning graduate students
- have the ability to write clearly for a general audience
- have the experience and skills for working in teams
- have experience with different levels of computer support for the same task
(e.g., use of a text editor vs. a word-processor for creating documents)
- have some programming experience and considerable experience using computers
- have at least one course in applied statistics
- have a course in psychology, preferably human information processing
(although this may be unrealistic)
Before discussing specific course proposals,
it is worthwhile to discuss general objectives for coursework in HCI.
The following are from the ACM SIGCHI Curriculum Development Group report.
By taking courses in HCI, we expect students to have gained an understanding of the following:
- the scope of issues affecting human-computer interaction
- the importance of the user interface to motivate the study of topics like HCI and user interfaces
- the impact of good and bad user interfaces
- the diversity of users and tasks (applications) and their impact on the design of user interfaces
- the limits of knowledge of individuals developing HCI systems
- the need to work with others, skilled in diverse areas such as software engineering, human factors, technical communication, statistics, graphic design, etc.
- cost/benefit trade-offs in HCI design
- different system development lifecycles including those particularly applicable to HCI systems (e.g., iterative design, implementation, evaluation, and prototyping)
- how HCI concerns can be incorporated into systems development lifecycles
- the need to evaluate system usability (e.g., someone will evaluate usability even if not the developer, and, in some cases, not evaluating constitutes professional misconduct)
- the existence of design, implementation, and evaluation tools for developers with diverse needs and technical expertise
- the information sources available on HCI
The few HCI Masters programs
tend to be run by concentrations university faculty who specialize in HCI.
Requiring a large specialized faculty would severely limit the availability of HCI programs.
But given the interdisciplinary nature of HCI,
it may not be necessary for a university to have many HCI faculty
to support an HCI program.
- Carnegie Mellon University:
This program is affiliated with the CMU HCI Institute,
and boasts one of the strongest HCI facilities in the world.
Their 12-month curriculum requires 4 HCI core courses, 2 HCI project courses, and 6 electives.
The list of electives spans 4 categories:
- University of Michigan:
This program is in the School of Information
with several world-class HCI faculty.
The Michigan program requires 12 credit hours in foundations,
and 21 hours in electives.
Practical engagements (e.g., internships) are required.
- Indiana University:
This program is in the School of Informatics,
and like SUNY Oswego, has limited HCI faculty,
and is planned for Fall 2001.
Their curriculum will require 4 core courses (only 1 of which is focused on HCI),
6 electives from at least two departments,
and produce a 6-credit project/thesis.
While the programs at CMU and Michigan are beyond the means of all but a handful
of universities in the world, there are similarities among all the programs:
- Few core courses, one of which is an HCI survey course
- Project work that extends beyond one term
- Many electives, electively drawn from many departments
(Indiana requires electives to be from more than one).
These departments include:
- Anthropology
- Business
- Communications
- Computer Science
- Education
- Graphic Design
- Industrial Design
- Law
- Library/Information Science
- Psychology
- Rhetoric
- Sociology
The breadth of the listings suggests that any student in an HCI program
would need a faculty advisor about which electives are most beneficial
to their degree.
The curriculum in the SUNY Oswego proposal
"Human-Computer Interaction and Systems Engineering"
was brief, with only course titles.
Semester 1
|
Semester 2
|
Semester 3
|
Statistics, Tools & Methods 1
|
Statistics, Tools & Methods 2
|
Studio Project (6 s.h.)
|
Human Computer Interaction
|
Cognitive Psychology
|
|
System Engineering
|
System Design
|
Elective
|
Display Design
|
Elective
|
Elective
|
SUNY Oswego HCI & Systems Engineering
>From discussions with SUNY Oswego faculty,
there appears to be some tension about requiring programming.
To me, it seems that requiring systems courses
would place the program exclusively in the computer science department,
exclusively for programmers.
Based on the
ACM SIGCHI CDG recommendations,
I recommend admitting students with diverse backgrounds because:
- it would allow programmers to work with non-programmers and vice versa,
which is more realistic and more enriching
- electives could provide a developer's focus,
a usability evaluator's focus, a graphic designer's focus, etc.
- it would increase the number of students in the program
Some of the course titles appear relevant but perhaps too focused for HCI:
- Statistics, Tools, & Methods:
The term "statistics" raises a concern of teaching psychology statistics,
appropriate for analysing data from controlled experiments,
but not as useful as more qualitative methods used in HCI practice.
- Cognitive Psychology:
This seems too narrow because it appears to preclude perceptual factors
(e.g., color, sound) and physical ergonomics (input devices, workstation design).
It may also include cognitive theories, which may have limited applicability.
I think a course on human factors of computing would be more appropriate.
- Display Design:
This could be generalized to interactive display design
to integrate the dynamic nature of displays, including
issues such as topology, navigation, history, etc.
- System Design/Engineering:
I think this could work well is the focus is on more abstract
design issues, such as data modelling, object-oriented design
(particularly design patterns).
It could be combined with information design and information architecture.
I think an understanding about how systems are built is important,
but instead of a systems engineering course, I think a course
on user interface architecture would serve well.
- Studio Project:
Although a double-credit course, I think it would be better
to spread a group project course over two terms.
One benefit would be that some students would be new hires
and some would be seasoned veterans and could apply
what they had learned to educate the first-timers.
In the following table, I have listed a program of possible required courses.
Rather than propose detailed syllabi,
I have listed books and/or resources to indicate the sort of topics
I think would make for a balanced program.
I have different degrees of confidence in different courses,
and have lower conviction about the non-project courses in the later semesters,
and for those, I have added some notes.
Semester 1
|
- General: Introduction to HCI
- Methods: Usability Evaluation
- Design: Interactive Display Design
- Human Factors: Human Factors in Computing
|
Semester 2
|
- General: Project I
- Methods: Analysis Methods
- Systems: User Interface Software Architecture
- This course would cover user interface tools and environments
that support user interface design patterns.
Ideally, it would include some high-level development tools.
- Collins 1995 Designing Object-Oriented User Interfaces
- Olsen 1998 Developing User Interfaces
- Gamma et al 1995 Design Patterns
- Elective
|
Semester 3
|
- General: Project II
- Design: Information Design/Architecture
- The main topic for this courses is the design of the entire
user interface and the information underlying it.
- Rosenfeld & Morville 1998 Information Architecture
- Henry 1998 User-Centered Information Design
- Elective
- Elective
|
The following electives seemed relevant (to varying degrees) to a Masters in HCI.
I have a preference for enforcing diversity by requiring
that students take electives from two departments out of their undergraduate major.
In the following list, which is a rough first cut,
I list courses from the SUNY Oswego 2000-2001 undergraduate catalog,
almost exclusively from 300-400 level courses,
although listing some courses that might be required for admission to the program.
Courses that seem especially relevant are bold and marked with a *.
Some that would be covered by required courses
are not allowed as electives.
- ANT: p.11
- 344: Language and Culture
- ART: p.15
- 307: Intro Graphic Design
- 308: Graphic Communications
- 309: Integrated Electronic Imagery*
- 312: Intro Illustration
- 407: Commercial Graphic Design
- 408: Commercial Graphic Design II
- 409: Digital Interface Multimedia*
- COM: p.28
- 301: Communication Theories
- 302: Message Criticism
- 310: Semantics
- 313: Writing for Public relations and Mass Media
- 315: Design and Layout for Media*
- 317: Public Relations Research
- 380: Communication Ethics
- CSC: p.32
- 241: Abstract Data Types and Programming Methodology
- 365: Data Structures and File Processing
- 366: Cognitive Science*
- 380: Software Engineering*
- 416: Artificial Intelligence Programming Languages
- 458: Computer System Analysis and Design
- 466: Artificial Intelligence and Heuristic Programming
- 470: Computer Graphics
- 480: Software Design*
- ISC: p.33
- 300: Current Problems in Information Science
- 320: Expert Systems and Knowledge Engineering
- 325: Multimedia/Hypermedia Design and Authoring*
- 329: Database Management Systems in Business
- 370: Human-Computer Interaction* disallowed
- 410: Decision Support Systems
- EWA: p.44
- 308: Creative Writing: Non-Fiction
- 408: Advanced Writing: Non-Fiction
- MAT: p.55
- 351: Intro to Forecasting
- 352: Statistical Quality and Process Control
- 358: Design and Analysis of Experiments
- PHL: p.65
- 321: Philosophy of Science
- PSY: p.75
- 280: Analysis of Psychological Data
- 290: Research Methods in Psychology
- 410: Human Factors* disallowed
- 401: Perception
- 405: Cognition
- 410: Individual Differences
- 480: Advanced Psychological Statistics
- 480: Tests and Measurements*
- 490: Research Seminar
- SOC: p.80
- 330: Qualitative Research Methods*
- MKT: p.91
- 370: Consumer Behavior
- 375: Marketing Research
- 390: Global Marketing
Some existing courses (ISC 370: HCI, and PSY 410: HF)
may be taught as is or adapted for joint graduate/senior service.
I recommend team teaching for many of the core courses,
especially initially when they are being developed.
This will ease the load on individual faculty
and provide multidisciplinary input into the courses.
Adjunct faculty might be brought in to teach some courses,
and these might be coordinated with (or be) industrial advisors.
Ideally, SUNY Oswego will be able to attract at least one HCI faculty
to focus on the HCI Masters program and coordinate the courses.
This faculty member may come from Computer Science,
but may come from Psychology, Cognitive Science or Engineering,
Human Factors, Industrial Engineering, or a Design discipline.
The difficulty of hiring good faculty in any field,
but particularly Computer Science,
requires creativity and flexibility.
Overall, I think the space and hardware requirements
for an HCI Masters can be minimal.
While visiting SUNY Oswego,
I was given tours of:
- Mahar Hall (Psychology)
- Wilber Hall (Graphic Arts)
- Snygg Hall (Computer Science)
- Penfield Library (Resources)
With the ubiquity of connections to the Internet,
almost any room with a network connection can serve as a design studio,
assuming the availability of portable computers
(which may be reasonable for most students focusing on HCI).
Preece 1994
provides some good ideas about design spaces
in the chapter about Supporting Design Teams.
Wilbur Hall had some open design areas,
and I would look to the design disciplines for design space ideas
because of their focus on studio/group experience.
More controlled environments may be needed for usability testing,
and some video equipment may be needed,
but a quiet room and video camera are probably adequate;
a dedicated usability lab
is almost certainly not needed.
The anechoic booths in the Psychology department are probably not suitable for usability testing.
Hardware requirements are minimal for many software packages.
Simple, but recent, PCs will allow most types of contemporary development
(e.g., Web, Applet, or Native applications).
I would anticipate little research by students
(e.g., like at Xerox PARC),
but students interested in high-end graphics workstations
might be able to make arrangements with faculty or local companies.
Specialized design or development software may be a major expensive,
but often there are inexpensive alternatives,
or educational discounts.
The Library seemed up to date, although with a limited budget.
Some books from a core list could be acquired by the library,
but faculty would need to assist students with getting materials
such as conference proceedings.
This could be done through private copies placed on reserve,
or through access to online materials
(e.g., the ACM Digital Library.
While talking with Cynthia Clabough in Wilber Hall,
I was looking at some health displays for handing out health information,
and she told me they were remakes and she told me about the one I was looking at:
"I didn't care for it, but people were picking up the pamphlets."
It was HCI in a nutshell.
I assumed that the remake was based on some sort of evaluation,
possibly by inspection or with real users,
and that the goal was that people would pick up the handouts
(and hopefully, these would be useful, usable, and used).
Equally important, there was an evaluation after the remake,
and the expert, I mean the professor, although not subjectively satisfied,
used the data and objectively concluded that the design worked.
To me, this is a good sign for an HCI program.
While SUNY Oswego is not a large school with extensive resources or faculty,
I think it has some key features that support an HCI program:
- Cognitive Science Cooperation:
While the HCI Masters may have little to do with Cognitive Science,
it is encouraging to see that different departments can collaborate
because it indicates that a multidisciplinary field like HCI
can expect similar cooperation.
- Graphic Arts Studio:
Few universities have the graphic arts presence available at SUNY Oswego.
The graphical nature of most user interfaces,
the design studio experience desired in HCI,
and the additional features of information design and evaluation,
are several strong signs for the success of the program.
- Software Design/Engineering and Mentoring:
The mentoring of students in project courses is a practice
that can be built on for HCI project courses.
- Human Factors:
The presence of highly motivated faculty ready to teach human factors
(based on industrial experience after a psychology degree)
is also a major component.
- The Whole is Greater than the Sum of its Parts:
The proposed courses
are similar those offered by
other HCI programs,
and the faculty are strong (although understaffed)
to teach them.
Although prospects for the program are promising,
I have observed some obstacles to overcome:
- Hiring HCI Faculty:
More faculty are needed, in part for program coordination and depth,
but also to alleviate the load on existing faculty.
Hiring HCI faculty will be difficult,
but looking outside of computer science may help.
- Getting Students:
There are geographical challenges to attracting students,
and the program may need to be designed to support distance learning
(in part), part time students, etc.
Industrial advisors might prove valuable.
- Computer Science or Broader:
I think the faculty need to decide if this will be an HCI Masters
program for software developers, or if the scope can be broader.
Some faculty may need to learn to accept the idea that different graduates
will leave have different skills because they came in with different skills
(perhaps they could have different concentrations).
Otherwise, I think the program may not attract a suitably
diverse group of students to create multidisciplinary teams.