Topic: interactive response time

user interface > Group: testing and evaluating user interfaces

testing and evaluating user interfaces
goals for a programming system
user interface
computer performance
ease of use
implementation of user interfaces
keyboard design
keystroke-level model for user interface evaluation
multi-processing for a user interface
programmer productivity
sense perception

Thesa - a thesaurus of ideas
topics f-m


Interactive computer systems must control response time to users. Basic response time should be practically instantaneous, as in display echo of typed characters. Command response time should be less than two seconds. Otherwise the user's attention will drift. Delays of longer than fifteen seconds break any feeling of interaction. Frequent system responses between 1/2 and 1 second should be consistent to support user interaction habits. (cbb 5/80)
Subtopic: responsive software up

Quote: responsive software lets users set their own work pace; keeps up with users, prioritizes feedback, signals busy state, and estimates time-to-complete [johnJ_2000]
Quote: responsive software provides instant feedback, estimates lengthy computations, supports multi-tasking, manages queued events, performs background tasks, and anticipates future requests [johnJ_2000]

Subtopic: time judgments up

Quote: most accurate time judgments between 0.6 and 0.8 seconds; detected 10% differences [millRB_1968]
Quote: several 'knees' is user's perception of response delay; e.g., 3 to 5 seconds [teitW3_1985]
Quote: when measured in information units, the performance capacity of the human motor system is relatively constant over a considerable range of task conditions; 10-12 bits/sec over much of the range from 1 to 10 bits/response [fittPM6_1954]

Subtopic: search time up

Quote: search for upper-left targets faster than lower-right targets; 2x to 4x; roughly linear for three column layout; lower middle targets slower [hornAJ9_2001]
Quote: 1/5 second search time for target-only layouts; independent of position [hornAJ9_2001]

Subtopic: productivity -- think time up

Quote: productivity increases substantially as response time decreases below a second or even half a second [bradJT5_1986]
Quote: productivity is better with 0.5 second response time instead of 3.0 seconds [thadAJ4_1981]
Quote: Visicalc was successful because it jumps; has a sense of liveliness, a world of linked action [shneB_1985]
Quote: Cedar goal of under five seconds for minor program changes; makes programming think-bound [teitW3_1985]

Subtopic: instant response, latency up

Quote: need <30ms latency for something to feel attached to your hand; need <50ms for keyboard echoing [gettJ1_2012]
Quote: lag degrades human performance on interactive systems; measurable at 75 ms and degraded at 225 ms; developed Fitts-based model [mackIS4_1993]
Quote: a responsive keyboard (click or force change) indicates responsiveness of the terminal [millRB_1968]
Quote: interactive systems should respond in less than 1/10 second to trivial requests; i.e., tasks without a sense of achievement or completion [bereG9_1982]

Subtopic: voice latency up

Quote: speed-of-light latency dominates voice-over-IP; need good access latency to stay below 150ms for voice [gettJ1_2012]

Subtopic: selection delay up

Quote: found eye gaze effective for displaying information relevant to an object; used 150-250 ms. dwell time [jacoRJ4_1990]

Subtopic: brief UI delays up

Quote: user interface delays should be brief, simple and standardized; e.g., user id confirmation [millRB_1968]
Quote: a user is captive to the terminal when waiting for a response; more than 15 seconds delay can be demoralizing [millRB_1968]

Subtopic: predictable response time up

Quote: if organize data for predictable working set size, get unpredictable response times in interactive systems; need random storage access [bereG9_1982]
Quote: increasing the variability of computer output significantly decreases user performance in interactive tasks [millLH6_1977]

Subtopic: rapid response up

Quote: response delays of more than 2 seconds should occur only after a task closure [millRB_1968]
Quote: a specification for response time should include acceptable tolerance values; should be below detectable range [millRB_1968]
Quote: good response time is critical for user satisfaction [rushA7_1986]

Subtopic: rapid link following up

Quote: point-and-click is 90% of a user's interaction with KMS; twice as fast as menu selection [akscRM7_1988a]
Quote: browsing requires immediate response; otherwise lose visual search and pattern-matching [bereG9_1982]
Quote: KMS users navigate by pointing at items linked to another frame; takes half a second [akscRM7_1988a]
Quote: fast link following is the most important parameter for Hypertext; goal is quarter second; twentieth second is too fast [akscRM7_1988a]
Quote: cognitive overhead in Hypertext is reduced by rapid access (KMS), short descriptions (Intermedia), and graphical browsers [conkJ9_1987]
Quote: Hypertext access time is important since readers need to look at nodes to determine relevancy [conkJ9_1987]

Subtopic: needless clicking up

Quote: Acme avoids needless clicking; no click-to-type, no pull-down menus, automatic window placement, context-dependent selection [pikeR1_1994]

Subtopic: tapping up

Quote: tapping the center of the screen is the most accurate, followed by the sides, the top or bottom, and the corners; corners require twice as much space as the center [hoobS5_2015]

Subtopic: rapid selection up

Quote: Alphaslider allows alphabetic selection from 10,000 titles in 13 seconds using 7 cm; current selection, coarse/fine thumbs, left/right arrows, initial letter index [ahlbC4_1994]

Subtopic: hour glass up

Quote: if an operation takes a long time in the Z editor, Z displays a line count periodically [woodSR6_1981]
QuoteRef: sammJE_1969 ;;690 types Wait on console while executing commands %type wait then digits to indicate progress

Related up

Group: goals for a programming system
Group: user interface
Topic: computer performance
Topic: ease of use
Topic: implementation of user interfaces
Topic: keyboard design
Topic: keystroke-level model for user interface evaluation
Topic: multi-processing for a user interface
Topic: programmer productivity
Topic: sense perception

Subtopics up

brief UI delays
hour glass
instant response, latency
needless clicking
predictable response time
productivity -- think time
rapid link following
rapid response
rapid selection
responsive software
search time
selection delay
time judgments
voice latency

Updated barberCB 7/04
Copyright © 2002-2023 by C.B. Barber
Thesa, Avev, and thid-... are trademarks of C.B. Barber