ArticlePDF Available

The importance of percent-done progress indicators for computer-human interfaces

Authors:

Abstract and Figures

A “percent-done progress indicator” is a graphical technique which allows the user to monitor the progress through the processing of a task. Progress indicators can be displayed on almost all types of output devices, and can be used with many different kinds of programs. Practical experience and formal experiments show that prograss indicators are an important and useful user-interface tool, and that they enhance the attractiveness and effectiveness of programs that incorporate them. This paper discusses why progress indicators are important. It includes the results of a formal experiment with progress indicators. One part of the experiment demonstrates that people prefer to have progress indicators. Another part attempted to replicate earlier findings to show that people prefer constant to variable response time in general, and then to show that this effect is reversed with progress indicators, but the results were not statistically significant. In fact, no significant preference for constant response time was shown, contrary to previously published results.
Content may be subject to copyright.
C H I ' 8 5 P R O C E E D I N G S APRIL 1985
The Importance of
Percent-Done Progress Indicators
for
Computer-Human Interfaces
Brad A. Myers
Department of Computer Science
University of Toronto
Toronto, Ontario, M5S 1A4
Canada
ABSTRACT
A "percent-done progress indicator" is a graphical tech-
nique which allows the user to monitor the progress
through the processing of a task. Progress indicators
can be displayed on almost all types of output devices,
and can be used with many different kinds of programs.
Practical experience and formal experiments show that
progress indicators are an important and useful user-
interface tool, and that they enhance the attractiveness
and effectiveness of programs that incorporate them.
This paper discusses why progress indicators are impor-
tant. It includes the results of a formal experiment
with progress indicators. One part of the experiment
demonstrates that people prefer to have progress indica-
tors. Another part attempted to replicate earlier
findings to show that people prefer constant to variable
response time in general, and then to show that this
effect is reversed with progress indicators, but the
results were not statistically significant. In fact, no
significant preference for constant response time was
shown, contrary to previously published results.
CR Categories and Subject Descriptors: D.2.2
[Software]: Tools and
Techniques-User interfaces;
1.3.6
[Computer Graphics]:
Methodologies and Techniques-
Interaction Techniques, Ergonomics.
General Terms: Experimentation, Human Factors.
Additional Key Words and Phrases: Progress Indicators,
Window Managers, User Interfaces.
Permission to copy without fee all or part of this material is granted
provided that the copies are not made or distributed for direct
commercial advantage, the ACM copyright notice and the title of the
publication and its date appear, and notice is given that copying is by
permission of the Association for Computing Machinery. To copy
otherwise, or to republish, requires a fee and/or specific permission.
© 1985 ACM0-89791-149-0/85/004/O011 $00.75
I.
Introduction.
Unfortunately, there will always be computer pro-
grams that cannot be executed instantaneously (fast
enough so the user does not notice a delay). Examples
of slow processes include compilers, text formatters, file
loading from floppy diskettes or other slow devices, file
transfers to remote machines or printers, and data base
processing. Even with supposedly interactive computer
systems that may have "easy-to-use" interfaces with
menus, icons or whatever, the user will still be faced
with periods when the computer has not finished pro-
cessing a request.
Percent-done progress indicators
are a technique for
graphically showing how much of a long task has been
completed. They operate like the giant thermometers
in charity drives and "fill up" from empty to full as pro-
gress is made (see Figure 1). Progress indicators give
the user enough information at a quick glance to esti-
mate how much of the task has been completed and
when the task will be finished. Many systems currently
present a "busy" picture, such as an hour-glass, clock, or
Buddha (for "patience'), to show that computation is in
progress, but since this is static, it does not indicate
how swiftly the program is progressing towards comple-
tion or whether the program has crashed or not.
Some systems, such as UNIX* and Accent"
(Rashid, 1981), support
multi-processing,
which means
that the computer can be performing more than one
task at the same time. When multi-processing is cou-
pled with a window management system, such as on the
BLIT (Pike, 1983) or PERQ* (Myers, 1984) then the
user
is encouraged to multi-process, i.e. run more than
one task at a time. For example, the user might
be
editing one file while having the system compile another
file in the background. Progress indicators can be used
in this case to show the progress of
each
process and
thereby keep the user informed about the state of the
entire environment.
*UNIX is a trademark of AT&T Bell Laboratories. Accent is a
trademark of Carnegie-Mellon University. PERQ is a trademark of
PERQ Systems Corporation.
11
C H I ' 85 PROCEEDINGS APRIL 1985
,.oo%
;o%
o%
Figure 1.
Percent-done thermometer that indicates approximately 70% com-
plete.
Figure 2.
Two icons from the Sapphire window manager (Myers, 1984), with
two progress indicators in each. The first icon shows 25% and 50%
complete. The second shows eight pieces of window and process
state information including the progress indicators at 75% and 99%
complete.
A typical implementation of progress indicators
will require that the application programs update them
explicitly. This means that if the program crashes, the
progress indicator will cease to be updated. Thus, pro-
gress indicators also tell the user if a program is still
running.
Progress indicators are usually presented graphi-
cally rather than shown as a numerical percentage. This
has a number of advantages: first, users can more
quickly and easily assimilate a graphical display than a
textual one (Myers, 1983) when an accurate value is not
required. Second, the graphical display implies that
only an approximate estimate of the time is available,
since exact times can only rarely be determined.
Finally, the graphical picture can be displayed in a
small space (as in Figure 2) without interfering with
other displays on the screen.
2. Implementing Progress Indicators.
Progress indicators can be displayed in a wide
variety of formats depending on the display device used,
but in every case, there must be some indication of the
percent of the entire task that has been completed. For
example, on a character terminal, a progress indicator
might appear as a series of asterisks along the bottom
of the screen, with completion signified by the asterisks
reaching the right margin (Figure 3). On a bit-map
display, such as those found on personal workstations,
progress indicators can be shown as a growing bar (Fig-
ure 3), an hour glass filling up, or a clock face with the
Figure 3.
A window from Sapphire showing a graphical progress indicator in
the title line, and a textual progress indicator using characters.
Figure 4.
Two other styles of progress display. The hand on the "clock" face
moves clockwise to completion, and the "sand" in the hourglass
moves from the top to the bottom.
hands moving (Figure 4). There should clearly be a
centralized routine to display the actual pictures so pro-
gress for all programs is uniform. In addition, there
may be auxiliary routines; for example, one might take
a file variable (as in UNIX or PASCAL) and show pro-
gress for the percent of the file read.
No matter how progress indicators are displayed,
programs will need to be able to calculate percentage
information for the indicators. This will clearly be easi-
est with algorithms that linearly process their input and
are then completed. Luckily, a fairly large number of
operations fall into this class. Examples include file
transfers, program loading, compilation, text process-
ing, etc. These account for a large proportion of the
long programs run on many systems. Unfortunately, all
of these may also have non-linear parts. For example, a
program to be compiled may refer to other programs
(such as files that are "imported" or "included") which
must also be processed. Also, the "piping" mechanism
in UNIX makes it difficult to tell how long the input
will be since it may come from another program. This
problem might be handled by insuring that all programs
in the pipeline are processing their input at approxi-
mately the same rate and basing progress on the origi-
nal data producer (either a file or a program creating
the data).
In programs that run multiple passes through data,
the progress indicator can be divided into sections for
each pass. Since progress is just an approximation any-
way, programs may be able to estimate how long they
will run based on heuristics or past experience. In addi-
tion, if a system supports a hierarchy of programs, for
example through command files or scripts, it may be
useful to present multiple progress indicators for the
same process. For example, the Sapphire window
manager (Myers, 1984) has two progress indicators in its
icons, one for the current program' and another for the
entire task (Figure 2).
12
C H I ' 8 5 P R O C E E D I N G S APRIL 1985
Figure 5.
The "busy bee" moves randomly around the screen to show that the
system is processing (PERQ, 1983).
Figure 6.
An icon from Sapphire with two progress indicators showing ran-
dom progress. While the application computes, the appropriate in-
dicator flickers.
For programs that simply cannot calculate how
long they will be running, a system can provide
random
progress.
This can be shown in a number of ways, such
as simply printing dots, moving around a "busy bee"
(Figure 5), or a constantly changing a pattern (Figure
6). This tells the user that the system is processing his
request and has not crashed, even though no informa-
tion is available to display the percentage completed. A
question, however, is whether having some programs
with percent-done progress indicators and some with
only random progress will be more annoying to users
than simply having no progress at all. Experience with
lOS (PERO, 1983) suggests that this is not the case.
Progress indicators are not a new idea. For exam-
ple, Spence (1976) reported that a graphical count-down
clock (Figure 4) was used to show the time left to com-
plete a request in a CAD-CAM application. They are
also used for file transfer in the Macterminal program
on the Apple Macintosh (Williams, 1984). For some
reason, however, progress indicators have only rarely
been used. Experience with the PERQ POS (PERQ,
1983) and Sapphire (Myers, 1984) systems, which have
thoroughly integrated progress indicators with their
user interfaces, have suggested that progress indicators
are very useful for a wide variety of applications.
There is clearly some cost, however, associated with
progress indicators in algorithm design and execution
time, so it is appropriate to try to determine if they are,
in fact, perceived as useful by users. This paper reports
the results of a formal experiment that shows that peo-
ple do, in fact, prefer to have progress indicators.
3. The Experiment.
3.1. Hypotheses.
This experiment was designed to test three
hypotheses. The first was simply that people preferred
systems with progress indicators. The second was that
the progress indicators would be more useful when the
response time of the system was variable rather than
constant. Earlier experiments, such as Miller (1977) and
practical experience (Carbonelle, 1968 and Weisberg,
1984) have shown that users prefer to have predictable,
constant response times rather than variable times, even
if the average time is shorter for the variable case. A
third hypothesis for this experiment was that this effect
could be reversed by having progress indicators.
3.2. Method.
A simplified computerized transportation manage-
ment system was prepared for this experiment which
performed simple pattern matching on a data base con-
taining about 100 travel entries. The task was similar to
the Miller (1977) experiment. Subjects first read an
instruction sheet explaining the task and the commands
available, and then they answered eight questions (see
Appendix A) using the computerized system, which
required about 14 queries to the data base. The ques-
tions were on a piece of paper and the answers were
written onto a separate answer sheet. The computer-
ized system ran on a PERO personal workstation
(Rosen, 1980) and was designed to be easy to use. To
make a query, the subjects filled out the form on the
screen (Figure 7), and then hit a key to have the system
match it against the data base. Before the results were
printed, however, there was a delay that was either a
constant 10 seconds, or randomly varied with a uniform
distribution from 1 to 17 seconds with an empirical
mean of 8.601 seconds. During this delay, a progress
indicator may have been shown. After completing all
of the queries, the subjects filled out a questionnaire to
gauge their feelings about the system. This featured a
"semantic differential" scale that attempted to measure
the subject's attitude towards the system. There were
10 items with a range of 1 (negative) to 9 (positive); for
example: "Anxious...Relaxed" or "Bored...Excited" (see
Appendix B). The aim was to measure various aspects
of the user's feelings. The dimensions were chosen
intuitively. The subject then repeated this process with
a different version of the system. The subjects were
divided into four groups to determine which versions
they used, as follows:
1: Constant time: First "Progress" then "No Progress"
2: Variable time: First "progre~" then "No Progress"
3: No Progress: First "Constant time" then "Variable"
4: Progress: First "Constant time" then "Variable"
Each group was further subdivided so that half got
one version first and the other half got that version
second. The system randomly assigned subjects to the
groups, with the constraint that there would be the
same number of people in each group.
After answering the second set of questions, the
subject answered the same questionnaire as before to
evaluate their opinion of the second version. Finally,
the subjects answered a different questionnaire that
asked the subjects to explicitly compare the two systems
and also asked some background information. All sub-
jects completed both versions and all questionnaires in
the same session, and the average times were about 5
minutes to read the instructions, 10 minutes for each
version, and 5 minutes for each questionnaire, for a
total of 40 minutes per subject.
13
C H I '85 PROCEEDINGS APRIL 1985
II
Commands:
-----]Oelete previous character. ~ Go to next field.
--~Oelete entry in this f'ielcl. ~ Erase all +'ields and start over
FI--NS--1Evaluate the form. ~-~Finished with all questions.
i Date:
Ronth: Jul
! Day:
I
! Destlnation:
Person Name:
People t:
Accom- z:
panying:
Progress:
Results:
1. Robert Frost to Kansas City on Jul ZZ Nith Rcla= Saith, Alan Sch~rtz.
Z. Larry
Nillar to
San Diecjo on Jul 4 Hith Rl~y Milson.
3. RI~x Nilson
to
Rtlanta on Jul 18 Hith Jane Fiero, Sis Ferar.
4. Robert Frost to Ne~ Orleans on Ju] ZI with Rrnold 5erkin.
5S~ Rl~y Hilson to Seattle on Jul |Z Hith David Simpson, Jane Fiero.
Larry Miller
to San
giL=cJO on Jul 5 with David Simpson.
?. Robert Frost to Rtlants on Jul 9 Nith David SimpSOn.
8. RI~F Milson to Rtlanta on Ju1 t Hith Rda= Ssith.
9. R¢lam Smith to Nm4 ,Orleans on Jul
15
Hith Sarah Smith.
19. Rdam Smith to Iliami on Jul 31 Hith David Simpson.
11. Rclam Smith to Des Roines on Jul 7.
IZ. Sil Farar
to
London on Jul
Z8.
Figure 7.
The screen for the forms-based query system used in the experiment. The subjects typed into
the various fields (e.g. "Month', "Person Name') and then hits the INS keyboard key to have the
query processed.
14
C H I ' 85 PROCEEDINGS APRIL 1985
1st
V:P C:P V:NP C:NP
NP:C
P:C NP:V P:V
2nd V:NP C:NP V:P C:P NP:V P:V NP:C P:C
Mean 1st 6.250 5.867 5.917 5.767 5233 6.067 6.367 6.216
Score 2nd 4.317 5.233 6.087 5.733 5.033 5.383 6.683 5.950
Table 1.
Unnormalized means of the scores on the semantic differential
(1=negative, 9=positive) on each of the 16 versions. Each column
represents one group of 6 subjects (each subject used two versions).
The top entry in each pair is the first version the subject used, and
the bottom is the second. Code: P=Progress indicator, NP=No pro-
gress, V=Variable time, C=Constant time.
Source
Subject
Var-Const
Prog-Not
First-See
V.C.*P.N.
Error
df Sum of Squares Mean Square F Value pr > F
47 10713.5729
227.9484
5.78 0.0001
1 65.3333 65.3333 1.66 02046
1 540.0208 540.0208 13.70 0.0006
1 4042604 404.2604 10.26 0.0025
1 94.0104 94.0104 2.39 0.1296
44 1733.8750 39.4063
Table 2.
Evaluation of significance of semantic differential scores.
1st V:P C:P V:NP C:NP NP:C P:C NP:V P:V
2nd V:NP C:NP V:P C:P NP:V P:V NF:C P:C
Mean 1st 4.16 7.83 7.33 7.83 6.33 8.33 6.33 6.50
Varia- 2nd 4.33 6.67 4.00 7.00 6.33 3.17 6.83 7.66
bility
Table 3.
Unnormalized means of the perceived variability of the response
time (1--very variable, 9=constant time) on each of the 16 versions.
Format is the same as Table 1. Code: P=Progress indicator,
NP=No progress, V=Variable time, C=Constant time.
Source
Subject
Var-Const
Prog-No
First-Sec
V.C.*P.N.
Error
df Sum of Squares Mean Square F Value pr > F
47 197.1667 5.507 1.45 0.1079
1 35.0208 35.0208 12.11 0.0011
1 7.5208 7.5208 2.60 0.1140
1 28.1667 28 3667 9.74 0.0032
1 35.0417 35.0417 12.12 0.0011
44 127.2500 2.8920
Table 4.
Evaluation of significance of the subjects' rating of the variability
perceived in their versions of the system.
3.3. Population.
Forty-eight subjects were tested, most of whom
were computer science graduate students, but about one
fifth were computer novices. All subjects were unpaid
volunteers.
3.4. Results.
The subjects did not have any trouble learning or
using the system. All versions were rated as easy to
use, based on the favorable comments from subjects and
the fact that there were few errors.
Table 1 gives the means for the subjects' score on
the semantic differential scale based on the different
versions. These data were fed into the SAS statistics
program (SAS, 1984) which generated the data for
Table 2. Table 2 demonstrates that the difference for
progress indicators versus no progress indicators is
highly significant (pr = 0.0006). Also, results from the
comparison questionnaire show that 86.1% percent of
the subjects liked progress indicators, and the mean rat-
ing for them was 2.94, where 1 means "Very Useful" and
9 means "Useless, Annoying." Other significant results
are that there was a substantial difference between sub-
jects, and the first version people used was rated higher
than the second.
An interesting result is that there is no statistically
significant preference for the version with constant
response time over the version with variable response
time (pr = 0~.046). Even when only the versions
without progress indicators are considered, the result is
still not significant (pr = 0.1497). Figure 8 shows the
average scores for variable and constant time for ver-
sions with progress and no progress.
"9ood"
6.0-
5.S:
S.e ~
S.?I
S.Sl
S.S I
?
"bad" I
0
I I
Variable Constant
Figure 8.
Graph of the mean scores on the semantic differential for progress
versus no progress and variable versus constant time. The sixteen
versions were summed into these four groups. It is interesting that
constant time is rated better (higher) than variable time when there
is no progress, but this affect is reversed when there is progress, as
hypothesized. This affect, unfortunately, is not statistically
significant.
Tables 3 and 4 show the results when the subjects'
evaluation of the variableness of the version is
evaluated. The significant result here is that there is a
high correlation between having a progress indicator
and correctly rating the variability.
3.S. Discussion of Experimental Results.
Clearly, this experiment strongly supports the
hypothesis that users prefer progress indicators. This
result is statistically significant at pr = 0.0006, which
means that there is only 6 chances out of 10,000 that
this effect would happen by random chance.
15
C H I '85 PROCEEDINGS APRIL 1985
Unfortunately, the last two hypotheses, that pro-
gress indicators would affect their feelings towards vari-
able response times, and that variable response times
would affect their feelings towards progress indicators,
were not supported. It is interesting to note, however,
that subjects prefered constant time (mean = 5.73) over
variable time (5.41) without progress bars, but prefered
variable (5.98) over constant (5.90) with progress bars as
hypothesized (see Figure 8). Since this difference is not
statistically significant, the experiment failed to dupli-
cate the results of earlier experiments such as Miller's
(1977) which said that subjects should favor constant
time over variable time, at least without progress indi-
cators.
By observing the subjects perform the experiment,
it was clear that when the progress indicator is present,
the subjects tended to watch it on the screen since they
had no other task to do. Without a progress indicator,
however, the subjects apparently got bored with the
screen and looked around the room or at the questions
or instruction sheet. When the answers appeared on the
screen, the subjects would notice this in their peripheral
vision, and then look up. This is supported by the data
in Tables 3 and 4 which show that the subjects rated the
variability of the constant and variable versions the
same without progress indicators, but there is a
significant difference when progress indicators were
present (correlation V.C. with P.N. at pr = 0.0011 in
Table 4).
This calls into question, therefore, the general
applicability of the earlier experiments; variable time is
not always perceived as worse than constant time. In
the Miller experiment, for example, the variability was
apparently in the rate at which characters were
displayed, which is an entirely different situation from
the one tested here. Clearly, if the degree of variability
is very low, the system will seem constant, and if it is
very high, e.g. 1 second to 1 hour, it will be unaccept-
able no matter what the mean is. An experiment to
investigate the range of variability that is acceptable,
with and without progress indicators, and under various
wait conditions, would be interesting. Another
approach would be to try to insure that the subjects
paid more attention to the screen, possibly by making
the test timed and having a very faint signal when
answers were ready, or by having the questions
displayed on the screen so there is no paper to look at.
Another interesting result is that subjects overall
had a lower opinion of the second version they used
than the first version which is statistically significant (pr
= 0.0025). This suggests that people got bored with the
system (due, no doubt, to the long response times) and
were annoyed at having to use it again. Since the
experiment controlled for this affect, however, it does
not bias the normalized results.
4. Interpretation of Advantages of Progress Indicators.
This section attempts to propose some explanations
for why people prefer the versions with progress indica-
tors. Since applications will typically not start showing
the progress indicators until they have parsed and
understood a command, progress indicators provide the
following important messages listed by Miller (1968):
the user knows (a) his request has been
listened to, (b) his request has been accepted,
(c) that an interpretation of his request has
been made, and (d) that the system is now
busy trying to provide him with an answer.
Progress bars are important for novice users since
they are likely to believe that everything on the com-
puter should operate quickly, and therefore is more
likely to panic (Foley, 1974) and think that the com-
puter has crashed if it does not provide feedback while
the computer is working. Although experts typically
have a feel for how long most tasks will take, they
should also benefit from progress indicators. Experts
will be more likely to run multiple tasks in parallel
since their time is valuable, but most people find it
difficult to keep track of what is happening when
multi-processing. Progress indicators help users plan
and monitor the various tasks so their time can be more
effectively used. An interesting experiment would be to
attempt to measure the affect of progress indicators in
a multi-processing environment.
Another reason people may prefer systems with
progress indicators is that they rarely like to sit idle and
waste time. Therefore, any "wait" time is annoying.
There are many examples of this effect in areas other
than computers, e.g. waiting "on hold" on the phone.
When people do not know how long the wait will be, it
is impossible for them to schedule a different task of
the appropriate duration, or even to relax effectively.
This tends to raise the level of tension while waiting for
completion. If there is an indication of progress, how-
ever, or if the user knows a priori how long the task
will take, then the time can be used in some productive
manner. This lowering of the users' anxiety is an
important benefit of progress indicators. Therefore, an
alternative to progress indicators might be an actual
number or an analog display of the actual time left until
the task is completed, if this can be estimated.
16
C H I ' 8 5 P R O C E E D I N G S APRIL 1985
5. Conclusion.
Percent done progress indicators appear to be an
important user interface tool that helps users in a
number of ways. They help novices feel better about
the system by showing that a command has been
accepted and the task is progressing successfully. They
are also useful for experienced users since they provide
enough information to allow them to estimate comple-
tion times and therefore plan their time more
effectively. This is especially important with multi-
processing systems with windows. The experimental
evidence presented here demonstrates that systems with
progress indicators are prefered by users. This indicates
that the benefits of progress indicators are probably
sufficient to warrant the extra cost in computation and
implementation required to include them in future sys-
tems.
Appendix A.
The questions that the subjects were asked to
answer using the query system in the experiment
included:
Who wanted to go to Fresno in December?
Who wanted to go to Santa Barbara on August 8?
How many requests for trips to Oakland are there in the
data base?
List the data of travel for requests that William Powers
made, where he requested Ned Maxwell to also travel.
If the people who accompanied Scott Derry to Oakland in
May, how many trips did each request for themselves in that
month?
Which of Walter Scdlak, Rosa Velaseo, or William Powers
booked the most trips?
Appendix B.
The following is the semantic differential scale
used in the questionnaire:
This version of the program made me feel:
Sad 1 2 3 4 5 6 7 8 9 Happy
Anxious 1 2 3 4 5 6 7 8 9 Relaxed
Impatient 1 2 3 4 5 6 7 8 9 Pstient
Annoyed 1 2 3 4 5 6 7 8 9 Calm
Tired 1 2 3 4 5 6 7 8 9 Energetic
Uncomfortable 1 2 3 4 5 6 7 8 9 Comfortable
Helpless 1 2 3 4 5 6 7 8 9 Powerful
Bored 1 2 3 4 5 6 7 8 9 Excited
Tense 1 2 3 4 S 6 7 8 9 At Ease
Confused 1 2 3 4 5 6 7 8 9 Confident
ACKNOWLEDGEMENTS
I would especially like to thank William Buxton for exten-
sive help in preparing this paper and the experiment. I would also
like to thank the many volunteers who participated in the experi-
ment. For help and support with this paper, I would like to thank
my wife, Bernita Myers, Neville Moray, Alain Fournier, and many
others at the University of Toronto and PERQ Systems Corpora-
tion. The research described in this paper was partially funded by
the National Science and Engineering Research Council (NSERC)
of Canada.
References
Carbonelle, Jaime, Elkind, Jerome I., and Nickerson,
Raymond S. (1968). On the Psychological Impor-
tance of Time in a Time Sharing System.
Human
Factors
I0(2), April 1968. 135-142.
Foley, James D. (1974). The Art of Natural Graphic
Man-Machine Conversation.
Proceedings of the
1EEE
62(4), April 1974. 462-471.
Miller, Lawrence H. (1977). A Study in Man-Machine
Interaction.
Proceedings of the National Computer
Conference 46.
AFIPS Press, 1977. 409-421.
Miller, Robert B. (1968). Response Time in Man-
Computer Conversational Transactions.
Proceed-
ings Fall Joint Computer Conference
33(part 1).
AFIPS Press, 1968. 267-277.
Myers, Brad A. (1983). Incense: A System for Display-
ing Data Structures.
Computer Graphics: SIG-
GRAPH '83 Conference Proceedings
17(3), July
1983. 115-125.
Myers, Brad A. (1984). The User Interface for Sapphire:
A Screen Allocation Package Providing Helpful
Icons and Rectangular Environments.
IEEE Com-
puter Graphics and Applications.
4(12), December
1984. 13-23.
PERQ POS Operating System Manual. (1983).
PERQ
System Software Reference Manual, POS Version
G3.
PERQ Systems Corporation, Pittsburgh, PA.
May 1983.
Pike, Rob. (1983). Graphics in Overlapping Bitmap
Layers.
ACM Transactions on Graphics.
2(2), April
1983. 135-160.
Rashid, R. and Robertson,
G. (1981).
Accent: A Com-
munication Oriented Network Operating System
Kernel.
Proceedings of the 8th Symposium on
Operating Systems Principles.
Asilomar, CA,
December 1981.64-75.
Rosen, Brian. (1980). PERQ: A Commercially Avail-
able Personal Scientific Computer.
IEEE Comp-
Con Digest.
Spring 1980.
SAS. (1984).
The Statistical Analysis System,
Release
82.4. SAS Institute Inc. SAS Circle, PO Box 8000,
Cary, N.C. 27511-8000.
Spence, Robert. (1976). Human Factors in Interactive
Graphics.
Computer Aided Design
S(I), January
1976.49-53.
Weisberg, David E. (1984). The Impact of Network Sys-
tem Architecture on CAD/CAM Productivity.
IEEE Computer Graphics and Applications 4(g),
August 1984. 36-40.
Williams, Gregg. (1984). The Apple Macintosh Com-
puter.
Byte Magazine.
9(2), February 1984. 30-54.
17
... Most prevalent front-end frameworks provide the default components to implement progress bar [53]. And users have a strong preference for progress indicators during long tasks [54]. We also recommend that the platform includes the progress bar in the same page of submission to create a one-stop solution for similar issues, such as the Privacy Dashboard mentioned in the previous section. ...
Preprint
Spoon Radio is a rapidly growing global audio streaming platform which currently operates in South Korea, the United States, Japan as well as the Middle East and North Africa. The platform believes that its commitment to user privacy is an important competitive factor. As such, it aims to not just comply with existing privacy regulations in regions where it operates today but to also ensure that it anticipates likely evolution to these regulations and of user expectations. In doing so, Spoon Radio wants to ensure it is well prepared to continue its expansion into new markets. As part of an effort to inform the evolution of its data practices, Spoon Radio reached out to the Privacy Engineering Program at CMU and sponsored a capstone project in which two master's students in the Program worked with Spoon Radio personnel over the course of the 2021 Fall Semester. The present report summarizes best practice recommendations that have emerged from this collaboration. These best practices are a combination of practices that are already implemented or in the process of being implemented by Spoon Radio today as well as more aspirational recommendations, which are expected to help inform Spoon Radio's practices in the future. In this report, best practice recommendations are organized around four stages of the data life cycle: data collection, data storage, data usage, and finally data destruction. A separate section is devoted to content moderation, an area where platforms such as Spoon Radio need to reconcile considerations such as promoting freedom of expression with the need to create a safe and respectful environment that complies with applicable laws and respects relevant cultural values.
Chapter
Full-text available
Batch processing reduces processing time in a business process at the expense of increasing waiting time. If this trade-off between processing and waiting time is not analyzed, batch processing can, over time, evolve into a source of waste in a business process. Therefore, it is valuable to analyze batch processing activities to identify waiting time wastes. Identifying and analyzing such wastes present the analyst with improvement opportunities that, if addressed, can improve the cycle time efficiency (CTE) of a business process. In this paper, we propose an approach that, given a process execution event log, (1) identifies batch processing activities, (2) analyzes their inefficiencies caused by different types of waiting times to provide analysts with information on how to improve batch processing activities. More specifically, we conceptualize different waiting times caused by batch processing patterns and identify improvement opportunities based on the impact of each waiting time type on the CTE. Finally, we demonstrate the applicability of our approach to a real-life event log.
Book
This book discusses human–computer interaction (HCI) which is a multidisciplinary field of study which aims at developing and implementing tools and techniques to attain an effective and efficient interaction between the humans (the users) and computers. In recent years, there is an increase of interest of HCI researchers and practitioners in the inclusion of gaze gestures which can greatly enhance the communication between the human user and the computer, as well as other more “physical” communication involving all what can be learned from movements of the human body, from face, hand, leg, foot, etc., to the whole body movement, even extending to the involvement of groups of agents, even society. These explicitly human-centric issues in the development, design, analysis, and implementation of the HCI systems are discussed in the book. A comprehensive state of the art is given complemented with original own proposals. As opposed to more traditional formal and IT based analyses, the discussion is here more focused on relevant research results from psychology and psychophysiology, and other soft, cognitive, etc., sciences. Remarks on the relevance of affective computing are also mentioned.
Chapter
What does the future hold for motion-based interaction methods? Will increasingly popular concepts and inventions, such as immersive virtual reality, hasten their already rapid development? Or will they be supplanted by other solutions, such as interfaces capable of reading brain activity directly, or those that recognize voice commands? Neither of these require movement. Each enables new perspectives for both users and technology, but also carry many risks. The development of technologies related to the recording of movement, such as immersive virtual reality, is also of high importance for other branches of science, including psychology. To date, no solution has facilitated the analysis of human behaviors by psychologists in such a detailed manner, nor in such natural environments.
Chapter
This chapter discusses the relationship between movement and mental and cognitive function, as well as the potential for it to be utilized by new technologies. The first section of the chapter presents the cerebral mechanisms responsible for associating physical with mental activity, and discusses examples of this influence on cognitive and emotional processes, as well as learning. The second section focuses on physical activity as an element of human interaction with a computer, including, but not limited to, so-called “exergames”. The subject of physical activity and cognitive functions is presented in the final section of the chapter from the perspective of immersive virtual reality technology—a tool which appears to be highly compelling. The potential of virtual reality stems from it being ideally suited to the study of the phenomenon of motion, and to its relationship with mental functioning. Immersive virtual reality is also a potentially effective motivator for increasing individuals’ physical activity, with a view to improving their mental functioning.
Chapter
This chapter focuses on eye movement from the perspective of human-computer interaction. The first section offers general information on the anatomy and physiology of the human eye, and outlines the key types of eye movement. It also presents the method of eye tracking and fields in which it can be applied. The chapter goes on to discuss issues pertaining to cognitive load, and how its intensity in relation to human-computer interaction and hypertext reading can be determined using oculomotor measurements. The final section is dedicated to the employment of eye tracking technology as a method of interaction. It presents methods that use the Gaussian function as a potential solution to the Midas touch problem. It also includes examples of solutions to registration errors during the use of head-mounted eye-trackers.
Chapter
This chapter presents concepts pertaining to movement in virtual reality (VR), with an emphasis on its prospective applications in psychological studies. Behavioral measurements, a key tool in such studies, are almost invariably conducted in unnatural and experimental settings, and entail considerable difficulty. These issues may be addressed by the application of VR. The chapter also illustrates examples of the technology’s use in studies of personality and differences between the sexes in psychological and psychiatric practice—for example, as a supporting tool in addiction treatment; and in more general applications that utilize VR in user movement tracking, such as the learning of dance movement patterns. The final section presents negative sensations experienced while using VR systems and discusses the mechanisms and causes of cybersickness.
Chapter
This chapter discusses the perception of movement during human–computer interaction. The phenomena discussed include, but are not limited to, perception of biological movement and deduced movement. The chapter also outlines the concept of mirror mechanisms. The later sections explain the potential effects of perceived motion on perception of the properties of an interface, such as its attractiveness, or its perceived performance. The chapter concludes with a summary of research on how movement affects the perception of avatars.
Chapter
This chapter discusses whole-body movement in the context of human–computer interaction. The first part focuses on the recognition and classification of body movements with the use of motion capture systems and video signal analysis. It also presents practical applications, such as automatic recognition of sign language and identification of individuals. The second half of the chapter reflects on the use of body movement as a method of interaction with computers and machines. Remote operation is discussed, including drone control. The chapter concludes with examples of new whole-body interaction paradigms.
Chapter
This chapter explores human–computer interaction involving the feet and legs. It begins with a concise description of the anatomy of the feet; next, it presents the elements of non-verbal communication that affect foot position and movement, as well as how the feet relate to aspects of human psychology. Both the anatomy of the feet and their connection to psychological functioning dictate how computer system designers involve them in interactions. The second half of the chapter outlines technological solutions that utilize the feet as a means of interaction with computers. It presents examples of indirect interfaces in which the feet are used to move various types of manipulator, such as pedals. It also presents direct interfaces based on sensors installed inside shoes—for instance, in soles or insoles. The final section discusses environmental sensors, such as sensing floors, which detect foot positions without the need to install special detectors or external controllers.
Article
Full-text available
One of the most important problems in the design and/or operation of a computer utility is to obtain dynamical characteristics that are acceptable and convenient to the on-line user. This paper is concerned with the problems of access to the computer utility, response time and its effect upon conversational use of the computer, and the effects of load on the system. Primary attention is placed upon response time; rather than a single measure, a set of response times should be measured in a given computer utility, in correspondence to the different types of operations requested. It is assumed that the psychological value of short response time stems from a subjective cost measure of the user's own time, largely influenced by the value of concurrent tasks being postponed. A measure of cost (to the individual and/or his organization) of the time-on-line required to perform a task might thus be derived. More subtle is the problem of the user's acceptability of given response times. This acceptability is a function of the service requested (e.g., length of computation), and variability with respect to expectations due both to uncertainty in the user's estimation and to variations in the response time originated by variable loads on the system. An effort should be made by computer-utility designers to include dynamic characteristics (such as prediction of loads and their effects) among their design specifications.
Article
The McIntosh microcomputer is introduced. Its characteristics are presented and a brief example is given to show how a program works. The foundation of the design is explained and four differences between the new machine and its predecessor, Lisa, are reviewed. The user-interface toolbox, the momory map, data sharing, languages, word-processing programs, graphics and several other features are described in detail.
Article
Many modern computer languages allow the programmer to define and use a variety of data types. Few programming systems, however, allow the programmer similar flexibility when displaying the data structures for debugging, monitoring and documenting programs. Incense is a working prototype system that allows the programmer to interactively investigate data structures in actual programs. The desired displays can be specified by the programmer or a default can be used. The default displays provided by Incense present the standard form for literals of the basic types, the actual names for scalar types, stacked boxes for records and arrays, and curved lines with arrowheads for pointers. In addition to displaying data structures, Incense also allows the user to select, move, erase and redimension the resulting displays. These interactions are provided in a uniform, natural manner using a pointing device (mouse) and keyboard.
Article
The literature concerning man-computer transactions abounds in controversy about the limits of "system response time" to a user's command or inquiry at a terminal. Two major semantic issues prohibit resolving this controversy. One issue centers around the question of "Response time to what?" The implication is that different human purposes and actions will have different acceptable or useful response times.
Conference Paper
The performance of users in man-machine interaction (MMI) is described in terms of a number of user- and machine-oriented parameters. The general linear model for experimental design is used as a model of the interaction. Performance measures are selected and a questionnaire developed to gauge user attitudes toward the man-machine system (MMS) and its environment. The interface parameters selected are hypothesized to have a significant effect on the performance and attitude measures. The effects of varying CRT display rates and output delays upon user performance and attitudes in a series of message retrieval tasks were evaluated experimentally. The results support the somewhat surprising conclusion that doubling the display rate from 1200 to 2400 baud produces no significant performance or attitude changes; increasing the variability of the output display rate produces both significantly decreased user performance and a poorer attitude towards system and interactive environment. The generally held notion that increasing output display rates is associated with better user performance is not supported.
Conference Paper
Accent is a communication oriented operating system kernel being built at Carnegie-Mellon University to support the distributed personal computing project, Spice, and the development of a fault-tolerant distributed sensor network (DSN). Accent is. built around a single, powerful abstraction of communication between processes, with all kernel functions, such as device access and virtual memory management accessible through messages and distributable throughout a network. In this paper, specific attention is given to system supplied facilities which support transparent network access and fault-tolerant behavior. Many of these facilities are already being provided under a modified version of VAX/UNIX. The Accent system itself is currently being implemented on the Three Rivers Corp. PERQ.
Article
Conventional wisdom has it that the range of analyses of which a computer-aided design system is capable is the primary factor in assessing its value in the design process. However, it is becoming clear that far more attention needs to be given to what are termed the human factors of a CAD system. For example, one of the principal objectives of such a system is to enhance the designer's insight into the product he/she is designing. Considerable potential for so doing is offered by the computer, but the application of this potential is often conspicuous by its absence. Similarly, the designer should be able to engage in a man – computer dialogue that is so designed that he/she is essentially unaware of the computer or the medium in which the dialogue is conducted. Again, this criterion is rarely met.The extent to which these two and other human factors requirements are satisfied will depend not only upon the skill of the cad system designer, but also on the medium in which the man-computer interaction takes place. Since interactive computer graphics offers considerable potential in this respect, it is useful to be aware of the considerations and techniques that are pertinent to this medium. It is this potential which is demonstrated briefly in the paper, mainly by means of illustrative examples. No excuse is offered for selecting most of them from a cad system with which the author is familiar, namely the minnie system for circuit design.
Article
High-level,predictable CAD/CAM performance can be realized with intelligent engineering workstations that are connected in a distributed/networked arrangement.