Please join us in cordially congratulating Dr. Steven Alter (professor emeritus, University of San Francisco), the 2021 AIS SIGSAND Outstanding Contribution Award (ASOCA) Winner. Steve Alter has made numerous and extensive contributions to the systems analysis and design community and The 2020-2022 AIS SIGSAND Executive Committee were unanimous in choosing him the 2021 ASOCA Award recipient.
Below is Steve’s autobiography (with links added by SIGSAND):
I received my PhD at MIT, taught at the Univ. of Southern California, and moved to San Francisco for family reasons.
I was lucky to work for 8 years in a firm that went from a small team of junior consultants to a successful manufacturing software startup.
I took a faculty position at the University of San Francisco while trying to figure out what to do next. One day a rep from a publishing company asked me whether I liked the IS texts that I was using. I said they were OK but actually would not have been significantly valuable to either the startup’s staff or its customers. That led to writing a lengthy IS textbook, each of whose successive editions tried to get closer to being of genuine value by seeing ISs in the context of work systems and organizations that they supported. A publishing merger led to a corporate decision that another edition would not be needed. Paul Gray, my former department head at USC, was starting CAIS around that time, and he asked me to help, especially by contributing articles and/or case studies.
I self-published a book on the Work System Method in 2006 after writing case studies and initial articles and conference papers related to the work system approach. My main goal was to clarify what I understood about the work system approach at that time, but I designed the book so that it could be used in teaching. Around 2010 I started to see that trying to publish academic papers about the work system approach would be a losing proposition unless I could claim it was a theory or could pretend convincingly that it was based on a theory, preferably a theory published previously in a respected journal or in a book by a philosopher. I published a JAIS 2013 paper on Work System Theory with substantial help from the SE and reviewers. Proposing a clear definition of WST turned out to be quite useful because it was now possible to say that WST consisted of 1) a definition of WS, 2) the work system framework, 3) the work system life cycle model. Identifying that core made it much easier to explain extensions of WST, i.e., ideas and frameworks related to service systems, workarounds, compliance and non-compliance, system interactions, design principles, etc.
I retired from getting paid by the Univ. of San Francisco in 2016 but continued going to conferences, doing projects with co-authors, and further extending ideas I had worked on for several decades. My current efforts focus on articulating an integrated view of systems in organizations called the Work System Perspective. This is a more expansive view of WST + WSM + WST extensions. The underlying assumption is that isolated papers about WST or WST extensions highlight separate topics but do not contribute enough toward a reasonably rigorous integrated approach that covers sociotechnical systems and totally automated systems. In effect that effort is like an attempt to produce and explain the Periodic Table of Elements instead of writing random articles about different elements.
Publications
Over the years, Steve published numerous highest quality research papers on systems analysis and design in top journals and conferences. We invite you to read his publications on GoogleScholar and ResearchGate.
Below, we provide three of his favorite papers on SAND:
I list the following three articles as favorites because each of them taught me something that I found important.
Markus, M.L., 1983. Power, politics, and MIS implementation. Communications of the ACM, 26(6), pp.430-444. Viewed at Power, politics, and MIS implementation (acm.org)
Although it was published decades ago, this article’s central message is relevant to SA&D today, i.e., that resistance to information systems may come from different types of sources including problems with the people, problems with the details of the system, and problems related to the interactions of people and systems. This implies that SA&D that lives up to its name needs to focus on more than rigorous specification of computerized artifacts.
Zur Muehlen, M. and Recker, J., 2008. How much language is enough? Theoretical and practical use of the business process modeling notation. CAiSE 2008, LNCS 5074, pp. 465–479, 2008. Viewed at 978-3-642-36926-1_35.pdf (springer.com)
That paper’s analysis of 120 BPMN diagrams found that less than 20% of BPMN was used regularly and that “the average model contains just 9 different BPMN constructs.” I saw its observations as a cautionary tale for thinking about different levels of detail in research and instructional materials related to my own efforts. The same issue exists today, where I am convinced that much of the value of the work system approach is simply in thinking about systems as work systems rather than IT systems, but where second and third level concepts and generalizations are quite useful in specific situations.
Alter, S., 2008. Service system fundamentals: Work system, value chain, and life cycle. IBM Systems Journal, 47(1), pp.71-85. Viewed at ibms-47-01-06 71..85 (researchgate.net)
I include one of my articles because working on it taught me to watch out for self-imposed limitations of simply drilling deeper and deeper into the existing ideas and approaches. Learning about IBM’s service science initiative had led me to wonder why and how service systems differed from work systems. This paper turned that question around a bit by assuming that service systems can be described as work systems but also that ideas related to service lead to ways of describing and analyzing work systems that would not have been imagined by just trying to drill deeper into previously existing work system ideas.
Interview with Steve
Steve kindly agreed to answer our questions, which we feel are especially relevant for anyone interested in systems analysis and design research.
Q1: Notable paper in SAND
Considering the long history of SAND publications, which one particular paper stands out for you, and why? What makes this paper special to SAND?
Steve Alter:
The following comments start with reflections on directions for extending one of my current favorite articles related to conceptual modeling (CM), go on to identify several possible titles for an ambitious PhD thesis, and then speculate about CM-related research topics that might be prominent a decade from now.
I especially like the way the following article opens a door to future research by focusing on moving beyond traditional assumptions and practices related to conceptual modeling (CM). It is great that a team of CM researchers managed to publish an MISQ article based on explicit reflections on limitations of traditional views of CM.
Recker, J.C., Lukyanenko, R., Jabbari Sabegh, M., Samuel, B. and Castellanos, A., 2021. From representation to mediation: a new agenda for conceptual modeling research in a digital world. MIS Quarterly: 45(1), pp.269-300.
I will summarize an even broader vision that might hint at directions for generating greater value from an expanded conception of CM but also might seem to stretch CM into an unrecognizable form. The broader vision is based on five premises:
Premise 1: Moving beyond linking ISs to representations. The MISQ article emphasizes limitations of seeing ISs as representations or fundamentally about representation. While the representation view inspired important research, many producers and consumers of CMs are concerned with topics that go far beyond representations. To be more valuable to many users, CMs related to ISs should be able to cover functions that ISs perform for work systems that they support. Those functions include providing, collecting, or transmitting information, enforcing rules for collecting and sharing information, controlling workflow sequence, controlling task execution, suggesting decisions, producing alarms when specific conditions occur, triggering automated activities, and performing automated activities. The MISQ article’s example of smart contracts makes a similar point (p. 283). A non-representational view of ISs says that ISs are work systems whose essential activities (deep structure) perform some combination of capturing, storing, retrieving, transmitting, deleting, manipulating, and/or displaying information. Many ISs exist to support other work systems that must be understood (and often modeled). As with other work systems, ISs may be sociotechnical (e.g., accountants performing accounting work to represent transactions and status information in a way that is both legal and beneficial for their employers) or totally automated (e.g., automatic generation of accounting reports based on stored transaction and status information and parameters provided by accountants). Notice that the difference between those two cases is not about a distinction between physical reality and digital reality.
Premise 2: Moving beyond the distinction between physical and digital reality. The MISQ article used this distinction effectively to motivate its explanation of its new view of CM scripts and grammars. That distinction is metaphorical rather than operational, as is clear from the nonobvious challenge of creating separate metamodels for descriptions of physical reality vs. digital reality. A footnote says that the distinction between physical reality and digital reality indicates the difference between lived versus computed human experiences (p. 270). Consider the online auctions that sell advertising slots in online magazines and newspapers in a fraction of a second when a new page is opened. The auctions are totally digital and therefore are not experienced by anyone, yet their results are “lived” through ad payments and through the serving and presentation of the ads. How would one decide whether they occur in a physical reality or digital reality, and more important, why would that actually matter to someone trying to analyze or design such a system? The same question applies to the smart contracts that are mentioned in the MISQ article. This entire area can be discussed in a much simpler way by identifying work systems that are relevant to a specific situation and decomposing those work systems to identify the relevant ISs, which may be sociotechnical (with human participants) or totally automated as in the accounting systems mentioned above. ISs are work systems that can be modeled using the same ideas that can be used for modeling any work system, regardless of whether it is sociotechnical or totally automated.
Premise 3: Relaxing restrictions related to script, grammar, method, and context. The MISQ paper says that both digital agents and human agents may be producers and consumers of CM scripts and that CM scripts may be grammatical or ungrammatical. A co-authored EMISAJ paper (Bork and Alter, 2020) addresses similar issues in a related way by proposing that modeling methods could provide more flexibility by relaxing traditional assumptions related to scripts, grammars, and methods. A broadly applicable modeling method for systems in organizations should be usable by business and IT stakeholders with diverse knowledge, beliefs, and roles. It should support different stakeholder purposes by including different techniques and different modeling languages, possibly based on different metamodels. It should maintain coherence across different purposes and metamodels by using a common modeling metaphor such as work system.
Premise 4: Integrating modeling, analysis, and design. Most people who use CM for real world problems (i.e., not just for teaching CM or for academic publications) should not see CM as an isolated exercise of documenting relationships between entities. Instead, CM should be integrated with analysis, design, and evaluation of real world systems and situations. Traditional separation between modeling, analysis, and design reflects specialist interests rather than needs of people trying to address real world problems. Some future form of CM might be viewed as part of the construction or self-organization of systems rather than as modeling prior to the construction of systems.
Premise 5: Making relevant background knowledge accessible. People who use CM for real world problems likely would benefit from some form of linkage between CM and accumulated background knowledge about systems that is broadly applicable and that was captured and organized outside of their own perceptions or analysis of a local situation. That knowledge about systems includes non-abstract knowledge, such as relevant examples, stories, and data, and abstract knowledge including concepts, generalizations (axioms, design principles, theories, frameworks, models, metamodels), and practices and methods. A greatly expanded version of CM might link conveniently with various forms of broadly applicable knowledge about systems. (In a CM world, such knowledge might be documented using relationships such as “potentially relevant to”.)
Those five premises include some ideas that are consistent with the MISQ article (e.g., limits of representation; relaxing assumptions about scripts and grammars) and some that imply extending CM in new directions, possibly at the risk of bending it out of shape (integrating modeling, analysis, and design; making relevant background knowledge about systems available). I hope those points are meaningful enough to encourage discussion and debate in the future.
Q2: Topic of a SAND Dissertation
If you were to do a Ph.D. on the topic of systems analysis and design today, what would the title of that dissertation be? What would the dissertation be about? Why?
Steve Alter:
1) “Integrated approach and suggested toolbox for modeling, analyzing, and designing work systems that might be sociotechnical or totally automated and that are subject to human agency and externalities.”
The point about work systems emphasizes the fact that a complete understanding of a situation requires modeling the relevant work system(s), not just an IS, software, or relevant information. Also, since people might or might not be participants in the situation being modeled and analyzed and might or might not follow rules or expectations, the analysis and design effort should anticipate the possibility that the system in practice may diverge from the system as designed.
The point about a toolbox is that modeling, analysis, and design require much more than just CMs in their current forms.
2) Another possible thesis topic is “Linking established knowledge about systems with modeling, analysis, and design of systems and organizations.”
A key challenge here is how to identify relevant knowledge and how to make it accessible. For example, specific items of knowledge might be identified in the form of knowledge objects (KOs) of different types such as characteristics (scalability, precision, degree of structure), performance variables (efficiency, accuracy, down time), phenomena (absorptive capacity, missing data, techno-stress), design principles (please customers, work efficiently, check inputs and outputs), theories (TAM, cognitive load theory, theory of constraints), etc., etc. Redundancy of the body of knowledge could be reduced by using inheritance across different special cases of work systems (e.g., WS in general, ISs, projects, IS development, open source IS development, open source IS development by nonprogrammers, etc.)
If that sounds too weird, then it may be useful to ask whether most people involved in modeling, analyzing, and designing systems have easy access to most of the generally available knowledge about systems that might help them contribute more effectively. This stream of research is unnecessary if they typically have most of the knowledge that they need (i.e., that they don’t really need any of the output of IS research). If they lack ready access to some valuable knowledge, perhaps there is a straightforward way to accumulate that knowledge and make it accessible to support their efforts.
Q3: SAND in the Future
When you think about SAND 10 years from now (i.e., around the year 2030), what do you think would be the most actively researched topic then? Why?
Steve Alter:
Here are some assumptions:
- The coming decade will see great progress in research under the broad umbrella of AI, including new combinations of machine learning, natural language processing, text mining, and process mining.
- “No code development” will advance greatly.
- New, easily combined deep structure building blocks may be created by applying some combination of text mining, process mining, and reverse engineering to existing ISs and to existing suites of commercial application software.
- New versions of tools in the general spirit of GPT-3 will be increasingly able to stitch together generalized versions of system-related code snippets and relationships.
- New ways of testing stitched together systems will be developed.
- Various aspects of those trends will merge into new forms of agile development.
If those assumptions are realized to some extent, then research about SA&D will focus on a series of topics that call for linkages to new types of CM:
- how to express maximally useful building blocks,
- how to consolidate those building blocks into trial versions of system descriptions,
- how to demonstrate those system descriptions to managers, users, and customers,
- how to realize those systems in totally automated target systems
- how to realize those systems in sociotechnical systems whose human participants may not fully understand the purpose, details, or impacts of those systems, may or may not like them, and may or may not perform their roles enthusiastically.
Thank you, Steve!
Once again, we would like to congratulate Steven Alter on receiving the 2021 ASOCA award and wish to extend our profound gratitude to him for his outstanding service and his intellectual and pedagogical contributions to SAND, and our society, broadly! Thank you, Steve!
