Ver 5 ZAIA only p 1 mdps workshop

Ver. 5 ZAIA only p. 1 mdps workshop Lyon June 2012 Design Strategies for Zero Adoption Impact Applications -The Ultimate in Ubiquity rev 1. 0 by Assoc. Prof. Karl Reed, FACS, FIE-Aust. , MSc, ARMIT Department of Computer Science & Computer Engineering, La Trobe University Visiting Researcher, Universita Degli Studi di Milano Dipartmento di tehnologie dell’Informazione

Ver. 5 ZAIA only p. 2 This Talk… mdps workshop Lyon June 2012 1. The case for Zero Adoption Impact Applications (ZAIA), the idea, its origins 2. Criteria for ZAIA applications 3. Simple examples of ZAIA systems in everyday life 4. Examples of ZAIA failures 5. An example of a potential ZAIA system, Wordtab 6. The research issue a design strategy for ZAIA applications 7. ZAIA criteria revisited-as philosophy, what else does it give us? 8. Conclusion And Now for something Completely Different!!!!

Ver. 5 ZAIA only p. 3 1. ZAIA, the idea, its origins mdps workshop Lyon June 2012 A. Examining possible future directions for IESE in the financial services sector showed v. large numbers of small businesses (in Germany) (EITO studies also show major concern about replacement of large-scale legacy systems. . ) Demographics of German Financial System *[Gfk**] Company Size No. Companies 1<9 91, 711 10<49 2, 474 50<199 1, 401 >200 1, 943 *Expect French, Italian Sector to have similar demography ** I have NEVER been able to find this reference since then! A warning to researchers

Ver. 5 1. ZAIA The only p. 4 mdps workshop Lyon case for Zero Adoption Impact Applications (ZAIA), the idea, its origins June 2012 B. This lead us to examine impediments to software adoption in this (and other small-business sectors). C. Impediments are… -acquisition cost -obvious -fitness for use -obvious -cost of adoption -we’d consider training in addition to acquisition cost -impact of adoption -not so obvious, need to consider disruption to patterns of business, could be THE show-stopper for 1 -4 person business D. So, what happens if we make minimization of the impact adoption our primary goal? ? Enter ZAIA

Ver. 5 ZAIA only p. 5 mdps workshop Lyon June 2012 Adoption Costs Due to impact of “learning” on ROI. . -simple exponential learning curve shows time required for equality of output for new and old processes in terms of productivity increase. c = time for learning to be 95% complete Productivity Increase Required Normalised Time to Breakeven. No. learning times c 50% 1 30% 1. 5 20% 2. 00 15% 2. 5 10% 3. 33 5% 7. 00 -doesn’t include lost opportunity cost or disruption to business patterns -implications for very small businesses are extreme. -what new technology gives a 50% productivity increase? ?

Ver. 5 ZAIA only p. 6 2. Criteria for ZAIA Systems mdps workshop Lyon June 2012 How do we minimize adoption costs? Model No. 1 -- New system is totally identical to existing systems, hence. . - No new learning required, - User only notices better performance. . - New systems actually invisible Examples. . - OS grade fixing minor bugs, - OS upgrade reducing resource utilization (very rare!) - Introduction of computer control of automobile functions -ABS, Engine Management

Ver. 5 ZAIA only p. 7 2. Criteria for ZAIA Systems mdps workshop Lyon June 2012 How do we minimize adoption costs? Model No. 2 -- New system proper functional superset of existing systems, hence, plus Model 1. . . - No new learning required, unless wanted by user - User only notices better performance. . - Added functionality can be “logical upgrade” of existing functionality, making it easier to learn Examples. . Could apply to any software if done correctly, New system identical to old, but has improved error recovery, for e. g. -Sewing machines -Refrigerators

Ver. 5 ZAIA only p. 8 2. Criteria for ZAIA Systems mdps workshop Lyon June 2012 How do we minimize adoption costs? Model No. 3 -- New system is totally identical to existing systems, but adds radical functionality that is easy to learn. . - New learning derived from functional need that motivated the new system. . - User obtains massive increase in productivity for very minor learning effort. Examples. . Adding extended back-end DB capability to a simple query system, Adding DB capability to tables in word. .

Ver. 5 ZAIA only p. 9 2. Criteria for ZAIA Systems mdps workshop Lyon June 2012 Some Issues on Learnability. . High levels of learning acceptable? ? Domains Of interest General Design Assumption Invalidates ZAIA? ? -very levels of skill are assumed in the users, Any skilled trade/clerical/professional No, apply except in special cases -very high levels of productivity gain are involved, Almost any domain If ZAIA not attainable, then may be ignored -very high levels of safety are implied by user competence Medical workers, air/transport, naval systems If ZAIA not attainable, then may be

Ver. 5 ZAIA only p. 10 3. Simple examples of ZAIA systems mdps workshop Lyon June 2012

Ver. 5 ZAIA only p. 11 3. Simple examples of ZAIA systems mdps workshop Lyon June 2012

Ver. 5 ZAIA only p. 12 3. Simple examples of ZAIA systems mdps workshop Lyon June 2012 The automatic formatting mechanism in Word that produces hanging tags. . -Proposed by the author 1981, and used as a student exercise for a toy office automation system…. -Introduced by MS ca 1994 in Word The word version does not work quite as well as the design proposed by Reed and students

Ver. 5 ZAIA only p. 13 4. Examples of ZAIA Failures mdps workshop Lyon June 2012 Apple OS 10 upgrades for Intel chips omit OS 9…. --Apps requiring OS 9 cannot be run, includes early versions of MS Office (OS 9 supported OS 7, totally seamlessly, early versions of OS 10 supported OS 9) -------Users need to buy new software, or migrate Apple 10. 7 (Lion) omits Rosetta…. --Apps running in power-PC code cannot be run. -------Users need to buy new software, or migrate. . Example of tool which no longer supported. . Eudora Adoption of SAP by major University… --1000 casual users disrupted MS Office not compatible over multiple releases. . --Old Word, Excel, PPT cannot be processed

Ver. 5 ZAIA only p. 14 4. Examples of ZAIA Failures mdps workshop Lyon June 2012 Reason to believe that examples of Post Adoption Trauma are related to failure to attempt address ZAIA issues

Ver. 5 ZAIA only p. 15 The Post Adoption Rollercoaster from [Rosenal, 2004] mdps workshop Lyon June 2012 What Rosenal DOES NOT say is that post adoption costs, (i. e. the total incremental spend on training, user support, impact recovery), are often many times the acquisition cost for the system!

Ver. 5 ZAIA only p. 16 mdps workshop Lyon June 2012 Consequences of ignoring ZAIA** Recent Australian experience is of a class of New Process Adoption Failure (usually IT) in which systems are completed to spec. , accepted, but seriously disrupt the organisations concerned. E Australian Customs Service cargo system [Booze Hamilton, 2006] , [stapleton, 2005] E Qantas Smartjet aircraft maintenance system [Woodhead, 2007] E SME’s post adoption problems not understood by the IT community [Barnes, 2006] E Negative reaction to new travel agent system* E The 20 yr old accounting system at a major university is replaced with the latest package. Significant disruption ensues, and substantial, unpredicted post-adoption *While collecting my airline ticket, the travel agent was complaining about a new system that was not working as well as the previous one. . And was less flexible. ** Ernesto Damiani , Maria G. . Fugini, Karl Reed On the Measurement and Prediction of The Evolution of Business Processes Under Repeated Use - Research Issues EQUITY 2007,

Ver. 5 ZAIA only p. 17 mdps workshop Lyon June 2012 Failure to Deal with Broad Design issues a ZAIA Failure? E E E Insufficient recognition of the fact that, where IT system are not simply “end-toend” automation without human involvement, there may be an irreconcilable mismatch (lack of “fit”) between the process being introduced and its new environment Where a non-IT dominated Business Process is concerned, this lack of “fit” can be even worse. Plus. . Insufficient allowance for the impact of human actors evolving a system with experience. (May be part of the PAT problem. . Human actors may not be able to see how the process will work when they are experienced with it? More later) Management often assumes that the existing process, and hence the human actors, are the problem, and ignore their potential contribution and role in an improved process. . an impediment to ZAIA design Major new processes (and their supporting IT systems) are sometimes made by senior management without proper reference to the IT and business unit domain experts, and for “political” reasons. [Larsen 1999][Besson 2001] *We are conscious that this issue is not unknown, but the fact that it keeps on happening means that it is not well understood or being taken seriously, at least in

Ver. 5 ZAIA only p. 18 mdps workshop Lyon June 2012 5. An example of a potential ZAIA system, Table-Base Using a table to create a hierarchy is natural. One does not keep creating new rows, one creates sub-rows often with a common root. There could be a number of “null” entries, where the user has aligned entries in different columns. Having the DB stored would allow sensible searches, and, searches to be made. The DB definition could be extracted from the table itself.

Ver. 5 ZAIA only p. 19 5. An example of a potential ZAIA system, Table-Base mdps workshop Lyon June 2012

Ver. 5 ZAIA only p. 20 mdps workshop Lyon 6. The research issue-a design strategy for ZAIA applications June 2012 Table V-Research Issues and Sources for ZAIA Methodology Development (Reed 2006) Overarching Factor Implications Preservation of -Develop new legacy Systems services using existing functionality -Reengineer to improve reliability, performance etc. -Migration to better costperformance platforms Research Area Detailed Topics (notes) -Identifying existing functionality within a legacy system and making it available via an API -(User centred SE -Users may be able to work IS community with previously known may deal withy functionality this already)) - Re-engineering - Performance engineering - Interpretive systems - ZAIA Issue

Ver. 5 ZAIA only p. 21 Table V-Research Issues and Sources for ZAIA Methodology Development(cont’d) mdps workshop Lyon June 2012 Overarching Factor Implications Research Area Detailed Topics (notes) ZAIA Issue High Adoptability -Business function oriented applications -Usability, HCI -Ease of learning -Domain specific processes -(may already be covered by IS researchers? ) -The fundamental ZAIA issue -Usability/ HCI -user centred design processes -HCI centred design processes -Business process capture -Dependability -SCS Low cost -Reuse -Application generation

Ver. 5 ZAIA Table V-Research Issues only p. 22 Development(cont’d) and Sources for ZAIA Methodology Overarching Factor Implications Research Area High usability -HCI -Business function oriented -User centred SE -Domain specific processes TTM -reuse -productivity issues -process -design efficiency Detailed Topics (notes) mdps workshop Lyon June 2012 ZAIA Issue

workshop Lyon 7. ZAIA criteria revisited-as philosophy, what else does mdps it June 2012 give us? Ver. 5 ZAIA only p. 23 - A Framework for the design of extremely low-adoption cost applications. - A Framework for identifying “invisible” applications for ubiquitous applications. - Enforcing a usercentric, domain aware approach to systems development and deployment Finally. . - A Research Agenda for the Software Engineering community. . The development of a ZAIA framework

Ver. 5 ZAIA only p. 24 mdps workshop Lyon June 2012 8. Conclusion - Failure to consider the ZAIA issue may be at the heart of major systems failure - ZAIA systems will be readily market able into well-established user communities which are otherwise saturated. And Now… Something Different