Increasing the yield of useful systems and services

1
Increasing the yield of useful systems and
services
2
Outline

• I. Exploring affordances
• II. Affordances in information systems
• Lotus Notes
• Mosaic
• III. The productivity paradox

3
Exploring affordances

• From Gibsons psychology (ala Norman) -- objects
  have affordances
• Affordances are the real and perceived properties
  of an object that suggest how the object could
  possibly be used
• A design succeeds to the extent that a resulting
  objects affordances match its intended uses

4
Thinking about doors
Unambiguous

• Long horizontal bar -- push

5
Thinking about doors
Unambiguous
Ambiguous

• Long horizontal bar -- push

• Short horizontal handle -- push or pull?

6
Systems with poor affordances

• VCRs -- programming timer recording
• UNIX -- cryptic commands, grep, awk, and
  vi
• Control rods on jet fighters -- upside down
  installation causes crashes

7
What leads to poor affordances?

• Make things invisible
• Be arbitrary
• Be inconsistent
• Make operations unintelligible
• Be impolite
• Make operations dangerous
• from Don Norman (1988) The Design of Everyday
  Things, page 179

8
Affordances in information systems

• Lotus Notes -- a groupware product
• Mosaic -- a graphical browser for the WWW

9
Lotus Notes Mixed messages

• Designed to share information in groups
• The interface does not afford this use
• records message posting behavior
• does NOT record message reading
• Manufacturing setting No evidence of managers
  interest -- declining message posting activity

10
Mosaic A success story

• Designed to facilitate navigation of the Web
• The interface nicely affords this use
• links are underlined
• clicking on a link connects to the desired URL
• Mosaic was adopted by millions within months of
  its release (and was the basis for the Web
  browser industry)

11
Understanding the productivity paradox

• Decline in productivity

12
Understanding the productivity paradox

• Decline in productivity
• Correlation with increased investment in
  computing

13
Is it mismeasurement?

• Mismeasurement of output
• BUT -- likely that we are underestimating labor
  input (due to unreported work)
• AND -- if we assume continuation of pre-1973
  patterns, there is an unmeasured shortfall of
  4000 per person
• does this seem likely? (p. 98)

14
Do we need more?

• Insufficient installed base
• BUT -- taking hardware and software together
  constitutes 12 of capital stock owned by US
  industries (Sichel, The computer revolution An
  economic perspective)
• AND -- railroads transformed the world at the
  same level of capital investment

15
Is it too early?

• Too soon to assess impact
• Electric motors didnt impact productivity until
  40 years after their introduction
• BUT -- we are long past Davids divide (50
  adoption)
• also, computers seem to perform about as well as
  systems they displace (i.e., same net return on
  capital)

16
Computers just arent very helpful

• 31 of systems built or installed are rejected as
  unfit for duty (Standish Group report)
• Hospital administrative staffs have grown as
  patient populations have declined (Strassman, The
  squandered computer)
• Systems are fixed investments that cant be
  dumped during economic downturns

17
The bottom line...
1,000.00
5,590.00
1,730.00
Depreciation
Applications
PC support
Network support
"futzing"
3,510.00
1,170.00
18
I futz, therefore I am

• 5.1 working hours per week
• Waiting for programs to run
• Waiting for help
• Double-checking formats
• Re-arranging files (finding files!)
• 100 billion in lost productivity per year
• SBT Accounting Systems

19
User-centered design

• Why do we need it?
• creeping featurism
• Software is a gasit expands to fill its
  container -- Nathan Myhrvold

20
The sad story of Windows 3.x

• Users couldnt manage overlapping windows
• More than 50 avoided running concurrent programs
• Most did not understand putting folders inside of
  folders
• Many couldnt double-click the mouse
• The average person took 10 minutes to open an
  application

21
Doing better?

• Maintain a focus on ease of use and efficiency
• Even technically savvy programmers rarely test
  interfaces on typical users
• Jakob Nielsens steps to better usability
• http//www.byte.com/bonus/117int.htm

22
The 10 steps

• 1. Provide status feedback
• 2. Match between the system and the real world
• 3. Give users control and freedom (provide an
  escape)
• 4. Be consistent
• 5. Prevent errors

23
The 10 steps

• 6. Make objects, actions, and options intuitive
• 7. Provide flexibility
• 8. Practice aesthetic and minimalist design
• 9. Help users recognize, diagnose, and recover
  from errors
• 10. Provide help that offers quick answers to
  specific questions

24
NYSE case

• 1993
• using circa 1987 technology
• volume of trading overwhelming systems
• needed change
• Approach
• six months of observation
• iterative testing of prototypes (30 cycles)
• the software engineers hated this approach

25
Results

• Two years to upgrade VS six years in previous
  revision
• Wireless handheld computers replaced paper cards
  and readers -- saving 1 million a year
• Order of magnitude drop in errors with doubling
  of capacity

26
A useful resource

• Much of this lecture is drawn from
• www.scientificamerican.com/0797issue/0797trends.ht
  ml

27
(No Transcript)