Com paring W eb Applications w ith Desktop Applications: An Em pirical Study Paul Pop paupo@ida.liu.se Department of Computer and Information Science Linköping University Sweden 1 of 13 May 26, 20 0 0
Motivation and Objective � Draw backs of desktop applications: development done on multiple platforms, have to downloaded before their use, administration and maintenance. � W eb applications: thousands implemented in recent years, used by millions of users. � Usability of w eb applications com pared w ith desktop applications. Com paring Web Applications with Desktop Applications: An Em pirical Study 2 of 13 Paul Pop May 26, 20 0 0
Definitions � Desktop application: WIMP, direct manipulation. � W eb application: runs on a server and presents itself through a web browser. Examples: web based email, bookmark managers, personal information managers, web calendars, online banking. Com paring Web Applications with Desktop Applications: An Em pirical Study 3 of 13 Paul Pop May 26, 20 0 0
Calendaring Application � Our specifications for an application: cost and size. � Calendaring applications : overview of a day, week, year; add, delete and move events; find free slots and events; reminders, meeting planners, sharing, “to do” lists. Com paring Web Applications with Desktop Applications: An Em pirical Study 4 of 13 Paul Pop May 26, 20 0 0
Microsoft Calendar Com paring Web Applications with Desktop Applications: An Em pirical Study 5 of 13 Paul Pop May 26, 20 0 0
Yahoo!Calendar Com paring Web Applications with Desktop Applications: An Em pirical Study 6 of 13 Paul Pop May 26, 20 0 0
Study Setup: Tasks � Training tasks. � Tasks: Go To From today, go to July 17. Add Add “Buy tickets” from 2pm, for 30 min. Move Move two consecutive events 1 hour. Undo Undo the previous move task. � Time per task in seconds, errors. � Small pilot study. � Record and Playback: ScreenCorder. Com paring Web Applications with Desktop Applications: An Em pirical Study 7 of 13 Paul Pop May 26, 20 0 0
Study Setup, Continued � Six subjects: five male and one female. � On average 48 minutes with the study. � Allowed to abort tasks. � Questionnaire: age from 25 to 27 years, median of 26.3 years, more than 5 years experience with computers, used computers more than 20 hours per week, three used the web applications, one used calendaring applications before. � Lim itations: subjects, statistical analysis. Com paring Web Applications with Desktop Applications: An Em pirical Study 8 of 13 Paul Pop May 26, 20 0 0
Average Tim e per Task 100 90 80 70 60 Desktop 50 Web 40 30 20 10 0 Go To Add Move Undo Com paring Web Applications with Desktop Applications: An Em pirical Study 9 of 13 Paul Pop May 26, 20 0 0
Error Rates Go To Add Move Undo MC YC MC YC MC YC MC YC OK 6 6 6 6 5 5 6 3 Error 0 0 0 0 1 1 0 2 Abort 0 0 0 0 0 0 0 1 Com paring Web Applications with Desktop Applications: An Em pirical Study 10 of 13 Paul Pop May 26, 20 0 0
Usability Problem s Com paring Web Applications with Desktop Applications: An Em pirical Study 11 of 13 Paul Pop May 26, 20 0 0
Discussion � Hypothesis (confirmed): the performance of users will be significantly reduced. � 3 factors: limited interaction mechanisms provided by web browsers, mismatch: user’s mental model and the application, delays from downloading the web pages. Com paring Web Applications with Desktop Applications: An Em pirical Study 12 of 13 Paul Pop May 26, 20 0 0
Conclusions and Future W ork � Users are tw ice as slow when using web applications: interaction mechanisms provided by the web browsers, lack of delimitation between browsers and web applications. � Future work: predicting the performance degradation, guidelines for web applications, m apping existing desktop applications to web. Com paring Web Applications with Desktop Applications: An Em pirical Study 13 of 13 Paul Pop May 26, 20 0 0
Recommend
More recommend
