Daniel Teichroew (1925-2003)
The 'father of CASE'
Biography and Bibliography of his Professional Life
This article is my personal presentation of a professional biography and bibliography for Professor Daniel Teichroew. I use the term 'professional' because it is not my aim to try to present a biography covering all aspects of his life. I have stepped outside his professional life in order to give some context for his early years. In terms of the bibliography, I have only included material external to the ISDOS project (see below) except for a small number of key references. He was known as Dan to his colleagues and friends, and I will use that throughout the article.
Daniel Teichroew - the phases of his life
Being born in 1925 (fifth of January in Manitoba, Canada), it was virtually inevitable that Dan would be affected by World War II. That is exactly what happened. He joined the Royal Canadian Air Force in 1942 (Teichroew 1944; Teichroew 1992), before he had completed high school, and proceeded to train firstly as a bomb aimer and then as an air navigator. His graduation from the Air Observer School was on 22 September 1944. Following that graduation, he went on to general reconnaissance until 1945. From 1945 onwards, his professional career started. Looking at his publications and records, I think it's safe to propose a set of phases for his professional life:
- 1945 to 1953 - higher education
- 1951 to 1952 - University of North Carolina
- 1952 to 1955 - National Bureau of Standards and University of California
- 1955 to 1957- National Cash Register Company
- 1957 to 1964 - Stanford University, Associate Professor then Professor of Management
- 1964 to 1968 - Case Institute of Technology, Professor and Head of Division of Organizational Sciences
- 1968 to 2003 - University of Michigan, Professor of Industrial and Operations Engineering, director of the ISDOS and PRISE projects, and various roles within ISDOS Inc.
Higher Education (1948 to 1953)
Soon after leaving the Royal Canadian Air Force, Dan commenced his higher education. First, he went to the University of Toronto. He obtained his BA in 1948, and an MA in 1949. He enrolled in the North Carolina State College (subsequently North Carolina State University) Department of Experimental Statistics, being awarded a PhD in 1953 (Teichroew 1953a; University 1953). Given the year of his PhD and the new emerging computer technology, it was very timely that his PhD was about the use of high speed computers and distribution sampling.
Early Professional Years (1951 to 1957) - Statistics
Dan started his professional career as many PhD students do, by taking on some supervision or teaching work, and helping the university with research contracts, commencing in 1951. His first publications (Teichroew 1951; Teichroew 1952) preceded the award of his PhD in 1953 (Teichroew 1953a).
Early career years were very firmly in the field of statistics. While completing his PhD, he established a working relationship with the National Bureau of Standards (NBS). The first publication I can find (1951) was in fact a NBS publication. Of course, that relationship started because of relationships between the University and NBS, and Dan was asked to work on various contracts while completing his PhD. This was very early experience of knowledge transfer work between a university and an outside organization.
By 1954 he was a prolific writer. Some National Bureau of Standards work became available as working papers for the NBS (Dixon and Teichroew 1953; Teichroew 1953b; Teichroew 1954b), and one more formal NBS Report was published in 1954 (Teichroew 1954a). He had manuscripts in preparation for "A Table of Millile Probability Points of the Incomplete Gamma Distribution", and, "Generalized Normalization Polynomials for the t and Incomplete Gamma Distribution". As far as I can ascertain from my archives, those manuscripts did not become NBS working papers, reports, or standards.
There was a series of statistical publications from 1955 to 1957 (Foster and Teichroew 1955; Teichroew 1955a; Teichroew 1955b; Teichroew 1955c; Teichroew 1956; Teichroew 1957; Teichroew et al. 1956; Teichroew and Sitgreaves 1957). One of these publications was a very short paper, but one that can be immensely useful to researchers because he made public an awareness of several as yet unpublished statistical tables (Teichroew 1955b). A consistent theme of his work was not only the provision of useful statistical tables based on a variety of statistical distribution, but also the issues related to the use of computers in the creation of such tables (rounding errors, and so forth).
His work in statistics surfaced again briefly in 1961 (Robichek and Teichroew 1961; Teichroew 1962; Teichroew and Sitgreaves 1961), in the period 1964-1967 (Robichek et al. 1965; Robichek et al. 1967; Teichroew 1965b; Teichroew et al. 1967a; Teichroew et al. 1964a; Teichroew et al. 1965a; Teichroew et al. 1965b; Teichroew et al. 1968). Statistics came full circle for him in the 1990s by way of support to other colleagues in the University of Michigan (Ali et al. 1991; Hanash et al. 1993; Hanash and Teichroew 1998; Oh et al. 1999) following his taking more of a back seat at ISDOS Inc.
During a brief period at the University of California, his publications show the early days of his interest in information systems, and more particularly his thoughts about languages. At the crossover in his career between the National Cash Register Company and the University of California was an intriguing contribution to a paper concerned with problems of machine language translation (Bull et al. 1955). Inspiration for the later work on PSL/PSA, which involved substantial simplification of statements of requirements, can be found in that paper:
"The conclusions to be drawn are that English (or Joyce) is exceedingly redundant, that it takes a lot of writing to get a little said, the quantity (linguistic mechanics) and quality (dictionary meaning and grammatical information) are not identical, and that, as a result, if we have to make a compromise, we will get a more useful translation by eliminating 80% of the volume instead of 96.7% of the meaning and grammatical information" (Bull et al. 1955) pp98-99.
This 1955 paper was published while otherwise he was in the middle of prolific statistical writing.
Middle professional years (1957 to 1964) - Stanford University - Management
Dan moved to Stanford University in 1957, initially as associate Professor then as full Professor of Management.
While at Stanford, Dan's original reflections on automating business systems were undergoing much development accompanied by several publications in the field (Teichroew 1958a; Teichroew 1958b; Teichroew 1959; Teichroew 1960; Teichroew 1963; Teichroew and Miller 1964).
For him, statistics, technology, and management were being brought closer together. This happened for him even at a 'micro' level with one IBM report (Hirsch et al. 1960) which is a fascinating study about errors in creating handwritten numbers inside boxes which will then be machine read. That report has relevance for students of motivation and indeed Hawthorne effects (although no Hawthorne effect studies are cited). He was involved in another IBM Report which applied statistical analysis to problems related to the electro-optical scanning of printed letters (Teichroew and Erdman 1963).
There was a period of time, roughly 1963-1965 when Dan was exploring a very futuristic concept of management being influenced by the display of information needed to underpin management decisions and control (Teichroew 1963; Teichroew 1965c). This thread of work immediately brings to mind the work being done by Stafford Beer with his thinking about display devices and control rooms, eventually tried out in Chile. In one sense, that very early work was at the beginning of a long thread in the industry leading to today's data mining and business intelligence work. Dan's work was futuristic in the sense that even today, although substantial amounts of data are now available to managers, that leap into display devices to supplement management information systems is still some way away.
The application of computers to support accounting learning was also explored at this time (Teichroew et al. 1963; Teichroew et al. 1964b).
It is not quite so evident from his papers which other writers had a key influence on his thinking. I had the benefit discussing this with him, and I am confident that important influences were from Canning (Canning 1956), and Young and Kent (Young and Kent 1958). Subsequently, Sammet also figured in his considerations of languages (Sammet 1969) but she had also picked up a more general theme at the time, of language to express requirements.
These influences are not surprising because not only were these writers key for the development of thinking about requirements for computer-based information systems to support business and management along with possible associated languages, but also Dan had the fortune of working with most of them in their respective organizations. Dan was a contemporary of Young and Kent at the National Cash Register Company. He was a contemporary of Canning at the University of California.
Statistics and management came together for Dan in the period 1964 to 1965 with two important publications. A substantial book, which must have been some time in preparation, provides a very helpful background and set of mathematical and statistical techniques that can be applied to a wide range of management problems. The textbook contains examples of the use of many techniques used today, accompanied by exercises to be done by the student. Answers to selected exercises are appended. This is a substantial book that would still provide today's readers with a very useful compendium of mathematical techniques to use for business and management (Teichroew 1964a). This was followed by a conference paper devoted to discussing the application of mathematics and statistics in business (Teichroew 1965a).
Early ISDOS Years - Case Western Reserve University (1964-1968)
This was the period when the firm foundations for the Information System Design and Optimization System (ISDOS) were built (Teichroew 1965c; Teichroew 1966; Teichroew et al. 1964b)
At the same time as constructing the ISDOS project, Dan was also strengthening his work on curricula for the use of computers in business, management, and the clarification of requirements (Teichroew 1967b). He was also building and strengthening his wider professional work through the ACM (Teichroew 1964b; Teichroew 1968). As noted below, these two professional threads would come together later.
A meta analysis of simulation languages was carried out and published (Teichroew et al. 1967b; Teichroew and Lubin 1966). The particular thrust for this line of work was to look at languages used in conjunction with simulation, rather than simulation per se.
During this period, the ISDOS project was born. The initial definition of the ISDOS project was created (Teichroew 1967a) and the first PSL/PSA manual was prepared (Steiger and Teichroew 1968). The project was started in August 1967 (Teichroew and Sibley 1969) at Case Western Reserve University.
Right from the beginning, the project was underpinned by multiple PhD projects, and these provided very important research foundations. Several PhDs were started during this time although completed after the ISDOS project moved to the University of Michigan (McCluskey 1969; Nunamaker 1969; Tremblay 1969).
Why did the term "Problem Statement Language" become one cornerstone of Dan's academic and professional life? Young and Kent's paper (Young and Kent 1958), is precisely about creating statements about problems. Parallel thinking by Dan about languages and machine translation between them, yields an obvious mission to seek a problem statement language to aid system building. From some of Dan's teaching notes, it is also clear that he used a paper by McGee (McGee 1963) in this mission. That paper talks explicitly about problems, languages, and moving languages from programming, to more abstract non-procedural statements. Sammet (Sammet 1969) had caught on to this emerging issue of problem statements, and has a section (pp 726-7) discussing "Problem-Describing Languages", so that issue existed explicitly as such by the mid- to late-1960s.
In architectural terms, the 'dream' or 'vision' behind the ISDOS project was the automation of system (computer-based) building as far as possible. PSL/PSA was the 'front-end' - the capture and analysis of requirements. Those early days saw the key ideas of the next stage represented as Systems Optimization and Design Algorithm (SODA) originated in one of the PhDs being supervised by Dan (Nunamaker 1969). My own papers (discussed below) within the ISDOS project bar one, were targeted at taking PSA outputs as inputs to generate systems using VAX Information Architecture products and were, therefore, targeted more at the SODA space than PSL/PSA.
Key ISDOS Years - University of Michigan (1968-1983)
The ISDOS project moved from Case Western Reserve University to the Department of Industrial and Operations Engineering at the University of Michigan in June 1968. What had been ISDOS Working Paper 1 at Case Western Reserve University was combined with a phase 1 report to become ISDOS Working Paper No. 1 at the University of Michigan (Teichroew and Sibley 1969). At the same time, what had been ISDOS Working Paper No. 11 at Case Western University became ISDOS Working Paper No. 11 at the University of Michigan. The project had been moved, retaining most of the original project documentation numbering scheme, which was then continued at the UoM.
There were other presentations consolidating the move to the UoM, such as that given to the Diebold Research Program (Teichroew 1969a) helping to reinforce links with the professional and management worlds, and others (Teichroew 1970a; Teichroew 1970b; Teichroew and Sibley 1970) that had the effect of making the concepts of ISDOS and PSL/PSA available to a broader range and number of people, and in some instances to a view of problem statement languages (Merten and Teichroew 1972) broader than just PSL/PSA.
The contribution of the ISDOS project to the University of Michigan was substantial in academic, impact, financial, and knowledge transfer terms. It must rank as being one of their largest research programs. I suspect that it also still ranks as the largest 'public' research projects in the field of requirements modeling and the automation of system building. There are some private research projects which are probably larger, such as those at IBM and TRW, and although those private projects have made very substantial contributions to the academic and public debates, aspects of those projects remain, for understandable reasons, proprietary.
The success of the ISDOS project at the University of Michigan was based on a very careful attention to the details of a balance between academic rigour, extensive engagement with, and involvement by, industry and government, combined with a practical infrastructure able to deal with some practical computer issues. Many universities can be helped by learning from the way the ISDOS project was set up and run.
Firstly, we can look at integration between the project's research and teaching. This is a thorny problem because many academics and their managers hold an ideological view that the involvement of academic staff in research will, ipso facto, improve the quality of teaching. The evidence for such a relationship is not convincing:
"Several meta-analyses of the literature on the research-teaching Nexus discredit the notion that faculty research productivity improves students' educational experience. Faculty research is not widely and effectively integrated into undergraduate courses." (Prince et al. 2007) (p. 290).
The same observation has been made when examining possible relationships between research and teaching for the individual members of staff:
"At the level of the individual member of staff, the simple models of staff who are heavily productive in research outputs being the most effective teachers, or that high productivity and research results in effective teaching, are clearly suspect." (Jenkins 2004) (p. 30).
Another paper, which would certainly appeal to Teichroew's statistical proclivities, also fails to identify any clear relationship between research and teaching, but does identify a range of variables that may influence teaching quality (Drennan and Beck 2001).
So, how did the ISDOS project manage to create a very successful link between project research and teaching being done by the researchers?
One important factor that can inhibit an interaction between teaching and research is lack of flexibility in terms of both curriculum and course descriptions. These have not been addressed in any of the statistical studies mentioned above. If the curriculum is designed with minimum or zero options when students take a course, and if the modules or components are defined very prescriptively, there is probably a lethal combination in place which would discourage interaction between research and teaching. It is essential that the curriculum is defined including electives that are relevant academically to the award being sought, and which provide an opportunity for academics to apply their research interests to a course. More to the point, a university's procedures for allowing changes to course structures should be flexible and responsive enough to evolve as research underpinning the course or module subject matter evolves. Similarly, the modules or components of courses need to have inbuilt flexibility. This can be achieved by having more generic module titles such as "Contemporary Issues", "Subject Fundamentals", or "Work-Based Learning".
Having inspected some of Dan's class preparation notes, it is clear that not only were the students informed about very contemporary approaches and issues in the subjects, but assignments and assessments were related directly to research being undertaken. Students knew they were working at the edge of the subject as it was at the time. This principle applied from undergraduate courses right through to PhDs. This was an exciting, vibrant learning community, where faculty and students knew their learning was up-to-date and they may even be helping to take things forward when they do their assignments.
The fundamental business and research models for the ISDOS project were simple, but very effective. The ISDOS project had clear aims about what issues it was researching. Right from the beginning, it was decided that the outputs of the project had to be useful to real organizations - including huge organizations. Not only was this a university research project, it was also established to produce things useful to other people. The foundation for the financing was setting up a 'club'. External organizations joined the club by becoming sponsors. This was a subscription model, not a bespoke reactive research model. Many sponsoring organizations went beyond their basic club membership fees and did other things, such as sponsor PhD students, or pay for activity beyond the planned ISDOS project activity. The project also produced software. Club members could use this software in their own organizations. There were innumerable instances of sponsors going on to write papers about their experiences and giving their suggestions to take the project forward. To illustrate the point, as a sponsor representative, I attended several conferences or review meetings in Ann Arbor and Europe. I authored or co-authored papers for those conferences (Darnton 1987; Darnton and Maddison 1985; Kang et al. 1984). I worked with a European team of consultants to roll out PSL/PSA throughout DEC Europe, and then worldwide. As part of that work, I commissioned ISDOS to assist with creating some tools that would automate the generation of functional specifications compliant with IEEE Standard 830 (Darnton 1987). This was very deep engagement between the project and one of its sponsors. The two-way knowledge transfer was of great benefit to both parties.
That field-testing was immensely valuable to the project. The project was the largest knowledge transfer project I have ever experienced (including recent experience with knowledge transfer projects), with the transfers being bi-directional. The University and the project sponsors knew what their strengths were and what they could contribute to each other. There was a 'club meeting', normally at least once a year. Most were in the US, and some in Europe. These club meetings started as ISDOS Review Meetings, and later were known as Information Systems Engineering Tools and Technology (ISETT) conferences. At those meetings, project academics gave papers, and sponsors gave papers.
So, not only did the ISDOS project tick the knowledge transfer box in an outstanding way, it also ticked the impact box in an outstanding way as far as usefulness to other organizations around the world was concerned. There is a very interesting discussion of this aspect of the project by Couger (pp3-5) in his later co-authored book about system development (Couger et al. 1982), discussing differences in approach between his 'school' and that of Dijkstra. The project also ticked another box: there was close cooperation between the academic and professional worlds, on an almost daily basis.
What about a narrower view of academic impact? After the move from Case Western Reserve University to the University of Michigan, the first few years were spent consolidating thinking in two primary ways. Firstly, the project was surveying not only languages that could help with requirements specifications, but also was grappling with the difficult technical issues of creating software that could implement PSL/PSA. The deeper research had been done in the PhDs completed in 1969 at Case Western Reserve University mentioned earlier, and other PhDs started at UoM after the ISDOS project was moved there. There were other publications exploring relevant issues ((Teichroew 1969b; Teichroew 1969c; Teichroew 1970a; Teichroew 1971a; Teichroew 1972a; Teichroew 1974a). In this period, the basic concept behind ISDOS was published in Datamation (Teichroew and Sayani 1971). There had been earlier surveys of approaches to stating requirements and Teichroew constructed what can be described as a meta-survey in which he extracted from seven different approaches, his view on the common characteristics that can be distilled from the prior surveys. The results of these deliberations were published in 1972 (Teichroew 1972b). As a meta-analysis, that 1972 paper has had ongoing academic impact. The 1971 Datamation article had an impact beyond the academic community, as it was also reaching a wider audience of senior people with management and executive responsibility in IT.
During this period 1968 to 1974, Dan was also active in the curriculum field as chairman of the ACM committee (1965-1971) that was designing an Information Systems curriculum (Teichroew 1971b). Dan contributed as Chairman of the ACM Curriculum Committee on Computer Education for Management to a further ACM Report that made recommendations for graduate information systems programs (Ashenhurst 1972). The first resulting book was published in 1974 (Couger and Knapp 1974). In that book, Dan reproduced some of his early publications related to the ISDOS concept and project (Teichroew 1974b; Teichroew 1974c; Teichroew and Sayani 1974). Dan moved away from that curriculum work to some degree following the 1974 book, but the curriculum work continued with a second volume published in 1982 (Couger et al. 1982). That second book contains 3 chapters in which Dan is an author (Teichroew and Hershey 1977; Teichroew et al. 1982a; Teichroew and Sayani 1971). Taken together, those two books contain valuable collections of key papers in the evolution of system development, including contributions by von Bertalanffy and Ackoff. They contain a collection of papers which would be very difficult to assemble individually today. Intriguingly, the key 1971 paper (Teichroew and Sayani 1971) was considered to be presenting a 4th generation approach to system development in 1974, but the same approach had become 5th generation by 1982!
As PSL/PSA is very much focussed at the initial stages of developing a new system, there is clear relevance of user requirements, and this is a topic that received some specific attention (Teichroew 1972a; Teichroew et al. 1975). At the 'front-end' of system development activity is also the problem of analysis, and this was the focus of the paper with Winters mentioned earlier (Teichroew and Winters 1976), with a strong emphasis on PSL/PSA followed by consideration of structured design issues (Teichroew and Gackowski 1977). Winters went on to complete his PhD (discussed below) which had important influence for the ISDOS project. Winters demonstrated the use of the ISDOS Language Definition Manager (also know to some as 'Meta') to create a langue for a Human Engineering Language for Information Systems (HELIOS).
1977 was a key year in terms of publication as that was when the paper that is probably the most cited paper about the ISDOS project was published (Teichroew and Hershey 1977). That paper goes well beyond problems of eliciting and expressing requirements because it talks about computer aided techniques that can assist, from requirements to system design. It discusses key problems with respect to documentation in terms of ambiguity, inconsistency, and incompleteness. The essence of the paper is that the production of requirements and design documentation really needs computer assistance because of the volume of information to be handled and to deal with problems of ambiguity, inconsistency, and incompleteness. Dan was a person who was very quick to spot whether a paper was a polemic or did indeed have some substance behind it. This key paper risks being a polemic because it asserts that computer aided documentation should be advantageous in terms of costs and benefits. Empirical work on precisely that question had not been carried out yet, and the paper recognizes this explicitly. Helpfully, a response to this weakness is provided in behavioural terms:
"the evidence that the PSL/PSA is worthwhile is that almost without exception the organizations which have seriously considered using it had decided to adopt it either with or without an evaluation. Furthermore, practically all organizations which started to use PSL/PSA are continuing their use (the exceptions have been caused by factors other than PSL/PSA itself) and in organizations which have adopted it, usage has increased" (ibid p47).
Interestingly, the paper, when discussing a range of other methods, makes the obvious observation that "to make the documentation more useful to human beings, many of these methods use a graphical language" (ibid p 43). PSL/PSA is a written language; it is not a graphical language. Reasons for adopting a linguistic approach rather than a graphical language involve profound issues such as readability and usefulness to human beings combined with the necessity to address matters of correctness, consistency, and completeness. The paper does not discuss this very important issue, even though very clear decisions had been taken to go down the linguistic root. The issue of graphical versus linguistic is one which still dogs the field of requirements analysis and representation. Automating completeness and consistency checking is the 'Achilles Heel' of all graphical techniques; for PSL/PSA it is very easy, and scalable to any size of system, precisely because it is linguistic with important 'constructor' properties (see below).
Looking at the extent to which this 1977 paper has been quoted in the academic and professional literature, there can be no doubt that this output of the project has had substantial academic and professional impact. At the time of writing this paper, Google Scholar was showing 462 citations including virtually all the key scholars and writers about requirements or system design tools. It is such an important paper in its field, that any writer who purports to present material about the evolution of analyzing requirements, or the automation of system building and does not cite that paper is at serious risk of not having done a proper literature review.
Following the 1977 paper, externally published output by Dan was low. This is not surprising as Dan's focus was on the ISDOS project itself. In this sense the list of references for this paper does not show the prolific generation of documents for the ISDOS project itself. My archives are not entirely complete. However, the highest Working Paper number I have is 672. The highest numbered technical memorandum is 571. There is some overlap in the numbering schemes, but at least 268 papers were presented at ISETT conferences. Therefore this period of time saw substantially more than 1,000 documents produced under Dan's direction, many of which bear his name, but are not formally referenced for this paper. This was prolific output.
External output during this period was focused primarily on the activities of the ISDOS project (Teichroew and Chikofsky 1980; Teichroew and Hershey 1982; Teichroew et al. 1982b; Teichroew and Sayani 1980; Teichroew et al. 1979), system development methodological issues more generally (Boydstun et al. 1980; Chaar et al. 1993a; Chaar et al. 1993b; Teichroew 1982; Teichroew 1993), and database management systems (Umar and Teichroew 1990). Dan was also an associate editor for a substantial volume about software technology, which appeared in 1979 (Wegner et al. 1979).
Over the years, a large number of organizations have been involved with the ISDOS project in ways such as being sponsors, contributing to ISETT conferences, or being in contact with the project for some substantive reasons. From the archives, this list includes: Advanced Technology Inc., ALMSA, ASTEC, ASWE, Aerospace Corporation, American Hospital Supply, AMK-Berlin, Aquidneck Data Corporation, Arabian American Oil Corporation, University of Arizona, ASEA, AT&T, Automation Industries Inc., Bank of America, Banca d'America e d'Italia, Banco Provincale Lombardo, Batelle, University of Belgrade, Bell Labs, Bendix Corp., Boeing, Boston University, University of Bradford, Brigham Young University, British Aerospace, British Railways, Bundesamt fur Wehrtechnik und Bescaffung, Burroughs Corp., CACI, California State College, University of California, Canada Defense Department, Carnegie-Mellon University, Case Western Reserve University, CDS Sistemi Italia, CCTA, CIA, Centrum voor Informatieverwerking, CERL, Chrysler, CIGER, CIISA, Citibank, Cleveland State University, University of Colorado, CSI, CSC, CDC, Connecticut General Life, Coopers & Lybrand, Agusta, Credito Commercial Bank, Deloitte, Haskins, & Sells, Technical University of Denmark, DEC, Diners Club, Dornier, Dynamics Research, EG&G, Technische Hogeschool Eindhoven, ESG, European Space Operations, Exxon, FAA, Fisher-Price, University of Florida, Ford Aerospace, Ford Motor Company, Fujitsu, GEC, General Dynamics, General Research, Georgia Institute of Technology, Grumman, GTE, Gulf Science & Technology Corp., Hadron Inc., Hahn-Meitner Institut, Harris Corp., Hasler AG., Hatfield Polytechnic, Hellenic Aerospace, University of Helsinki, Hitachi, Holiday Inn, Honeywell, University of Houston, Hughes Aircraft, IBM, Namur University, Intel, IPACRI, IPL, Israel Aircraft Industries, ITT, JPL NASA, Eastman Kodak, Korea Institute of Electronics Technology, Kyoto University, Litton Systems, Ericsson, University of Maine, Marconi, Martin Marietta, MBB Flugzeuge, McDonnell Douglas, McMaster University, University of Melbourne, Metropolitan Life, AFDSDC/SDBA, University of Minnesota, Mitre Corporation, Mitsubishi Electric, Mobil Oil, University of Montana, University of Natal, NATO, NCR, NCE, NEC, New Mexico State University, City University of New York, Northwestern University, University of Oulu, Pacific Gas and Electric, Pacific Telephone, Banca Antoniana de Padova, Perkin-Elmer, Petroleos Mexicanos, Phillips Petroleum, PTT-Netherlands, Purdue University, Raytheon, Rockwell, RAE, University of Saskatchewan, Scott Paper, Shape Technical Centre, Siemens, Southern Illinois University, Social Security Administration, Softech, University of South Africa, Southern California Edison, Sperry Univac, SRI International, Standard Oil, Standford University, Hochschule St. Gallen, Universitat Stuttgart, Syslab, Systems Designers, Tandem, Tektronix, University of Texas, Texas Instruments, 3M, Toshina, TRT, TRW, Upjohn Company, US Army, US Air Force, US Navy, Wang Laboratories -- just to name a few. This list is not complete, but illustrates the extensiveness of interest in ISDOS and/or PSL/PSA.
The ISDOS project was founded on some fundamental research about the construction of languages for defining systems. Clearly the project devoted most of its effort to PSL for capturing information about requirements, and subsequent analysis using PSA. The principal software manifestation of that fundamental research was known as the Language Definition Manager (LDM - called 'Meta' in some documents, and ISLDS - Information System Language Definition System - in others). This is explained in papers published from 1980 (Teichroew et al. 1981; Teichroew et al. 1980a; Teichroew et al. 1980b), and hinted at on page 44 of the 1977 paper (Teichroew and Hershey 1977) where it states " The Problem Statement Language is based first on a model of a general system, and secondly on the specialization of the model to a particular class of systems, namely Information Systems". What lies behind this is quite profound for at least two reasons. Firstly, it is a mistake to think that PSL/PSA is primarily a tool for defining computer-based systems. It is certainly a computer-based tool, but it can be used to define systems that are not computer-based. Secondly, LDM was used in the ISDOS project to create different languages. For example, the environment was used to create a language for CCITT (Block et al. 1983; Teichroew and Kang 1984). Outside ISDOS, LDM has been used to create a variety of different languages such as: Business Process Language; Information Architecture Language; Competence Modelling Language; and, Business Analysis Language. There is a paper discussing some of the results of using LDM to create new languages (Darnton 2002). There is earlier reference to the PhD by Winters who produced a different language. There are numerous other examples. This puts LDM into a category of being a meta-meta-modelling language. It remains a mystery why that fundamental work has not yet been published properly. However, further discussion of this is beyond the scope of this paper. LDM as incorporated into the ISDOS project was derived from Yamamoto's PhD thesis (Yamamoto 1981). Another PhD completed in the same year (Spewak 1981) explored developing a language from the same foundations, for the purpose of simulation.
Dan was also helping to supervise a PhD by Winters (Winters 1981). Winters not only applied exactly the same approach as set out by Yamamoto in his thesis, and presented in a conference paper (Teichroew et al. 1980b), but also he has a fascinating discussion (pp44-49) of Carnap (Carnap 1967) who must be the originator of object, property, relationship modelling (subsequently instantiated as entity, relationship, attribute modelling by other writers),along with the fundamental database ideas of one-to-one, one-to-many, and many-to-many relationships. That is another topic for another paper. It is sobering to reflect that such a modelling approach had been delineated by 1928 (Carnap 1928) although not immediately accessible to the English-speaking world. Object, property, relationship (OPR) modelling pre-dates entity, attribute, relationship (ERA) modelling and that as far as I can identify, the real origin of those is Carnap. Terms such as 'object', 'property', 'relation', 'one-one', 'one-many', are all there in Carnap. If this is correct, then by the 'bringing to market' of OPR and ERA modelling, the ISDOS project had a profound impact on subsequent academic and professional thinking about domains such as system building and data modelling. This puts ERA and OPR modelling very firmly as derivatives of the 'Vienna Circle' school of philosophy. The ISDOS project certainly had impact!
The definition and construction of PSL/PSA were undoubtedly created with Carnap's idea of a constructor in mind. In other words, while defining a language it is also necessary to define a corresponding 'constructor' in order to be able to verify completeness and consistency at any stage of constructing a model of part of the real world. PSL/PSA was a language and associated constructor combination. More fundamental than that, the combination of the ISDOS Language Definition Manager and System Encyclopedia Manager meant that when constructing any language, not just PSL, a corresponding constructor would also be generated. This is powerful stuff, and it is precisely where PSL/PSA competitors such as UML fall down; there is no 'constructor' defined for UML which means it is virtually impossible to carry out any kind of automated checking of UML outputs for completeness and consistency. If this is correct, PSL/PSA remains the strongest contender yet produced, to support the front-end of computer-assisted system building. In practice, it is this characteristic of PSL/PSA which has helped a great deal in requirements defining projects; when people are expressing their requirements, discussions take place at different levels and with differing detail. I had defined a set of database queries able to use the constructor properties of PSL/PSA to give periodic reports about what still remains to be clarified (Darnton 1987). I obtained the agreement of DEC to commission the implementation of those checks from ISDOS Inc. They were immensely useful. That work and resulting tools is known as FSGEN.
Transitioning ISDOS Years - ISDOS Inc., and PRISE (1983-1991)
Around 1982, the idea was mooted to spin off the ISDOS project into a separate privately held company. The reasons for doing so are completely understandable for any university-based project that has become successful and large. ISDOS was no exception. It had a large sponsor base, it was developing complex software, and it had multiple relationships, some very complex, with a wide range of external organizations. There was increasing demand from organizations deploying PSL/PSA for formal support arrangements. The University supported the idea for reasons including potential benefits for the general public and the University of Michigan, more effective dissemination of new technology in the public interest, generation of royalty revenue for the University community, and the ability to provide support for sponsoring organizations. Taking the project forward really needed an injection of capital. Every university dreams about generating new intellectual property that can be used as the basis for a stream of royalty income. It is not normally the job of a University to provide capital for a venture which is beginning to look as though it needs a proper commercial footing. Therefore, at the time, this looked like an obvious way forward.
Unfortunately, as things turned out, 1983 saw the beginning of a very turbulent period of time, finally resulting in the disappearance of PSL/PSA from the market and the research world. The University's aim of furthering the public interest was eventually frustrated.
On 14 January 1983, papers were filed to incorporate Information Systems Development Inc., trading under an assumed name of ISDOS Inc. Shortly afterwards (11th February 1983), the University of Michigan entered into agreements with ISDOS Inc., purportedly giving ISDOS Inc exclusive rights for the further development and marketing of PSL/PSA etc. From that point onwards, the original university project continued under the acronym PRISE (Program for Research in Information Systems Engineering). For what it needed to achieve, ISDOS Inc remained seriously undercapitalized.
Dan was now involved in both ISDOS Inc and PRISE. Academic output diminished, although there was some external output focused more generally on methodologies (Teichroew 1985) and a substantial book on the topic (Teichroew and Dávid 1985). However, the ISETT conferences continued, and provided a valuable platform for project sponsors and ISDOS Inc customers. Therefore most project output is properly characterized more appropriately as knowledge transfer rather than academic.
By 1985 there had been changes to the management structure of ISDOS Inc, but the problems associated with under-capitalization continued. Early in 1986, the company was quietly exploring the possibility of finding investors by talking with some known interested parties. On 11 June 1986 Information Systems Development Inc (trading as ISDOS Inc) changed name to meta [sic] Systems Ltd., with Richard J Welke as President and CEO. Dan continued as Chairman of the Board. Substantial code re-engineering work was done by Meta Systems Ltd., and a revised product set was put in place. A new service for reverse engineering was made available and made some progress. As well as changes to the product and service set, there were substantial changes in terms of management, organization, and marketing. Significant effort was put in to convert ISDOS sponsors to be customers of Meta Systems Ltd. The business of Meta-Systems Ltd was stated as:
"specifying and delivering (meta) systems which, in turn, aid organizations in producing more highly organized and effective end-user systems in an efficient manner. The change also marks the completion of our transition from a university-based research project to a professionally managed commercial organization" (Welke 1986).
Notwithstanding the substantial changes following the evolution from ISDOS Inc to Meta Systems Ltd., the company was dogged by two difficult problems. Firstly, the company remained seriously undercapitalized with respect to its ambitions. Secondly, there was a problem in terms of identifying the extent to which the University of Michigan really was able to have given exclusive rights to ISDOS Inc in the first place. Many sponsors had obtained software and access to project academic outputs, therefore there was a real question whether it was possible for UoM to convey exclusive rights to any single organization. This debate even extended to suggestions, and some legal argument, that some of the project software had effectively been placed in the public domain.
By 1989, Meta Systems Ltd and LBMS were in talks with a view to LBMS taking over Meta Systems Ltd. LBMS was a UK company and in 1989 it had completed successfully work helping the UK government to define a standard method for developing information systems known as SSADM. The takeover happened and Meta Systems Ltd in Michigan was dissolved on 19 September 1991. At the time of the takeover, LBMS had their own tools to support system development. I have no basis on which to know the real intentions behind the takeover, but in the light of what actually happened, I can only presume that the commitment by LBMS was to its own tools, and not to PSL/PSA. Effectively, PSL/PSA died and was not available to the market. I understand that LBMS used some of the software acquired from Meta Systems Ltd in its own product set, but at a deeper level and not evident to the users of other LBMS products. Following a major UK project that needed the use of PSL/PSA, LBMS transferred rights in the product to me.
When I asked Dan what was the largest system he was aware of which had been modelled using PSL/PSA he replied 14,000 objects with 65,000 relationships. The ISDOS project certainly solved some key system development problems and produced an (almost) industry-strength product.
To all intents and purposes, academic output related to ISDOS and PSL/PSA ceased (although academic work didn't). As mentioned above, Dan's academic output reverted to being more statistical, and some work on CASE mentioned below.
A helpful discussion, including comments about the history of PSL/PSA was published in 1990 by the co-author (with Dan) of the key 1971 Datamation article (Sayani 1990). Sayani has a very interesting discussion of issues related to the deployment of PSL/PSA in organizations.
The 1992 J-D Warnier Prize Winner
In 1992 Dan was 'The 1992 J.D. Warnier Prize Winner' for his outstanding contribution to the theory and practice of information science. He was cited as 'The Father of CASE'. The members of the selection committee were Barry Boehm, Gordon Davis, and Tom DeMarco. He prepared an autobiographical sketch for the award, and an edited version accompanied the prize (Teichroew 1992). He explains the reasons for becoming interested in statistics. His introduction to computers happened while studying at Toronto University. It was while at the National Cash Register Company that he became interested in problems of expressing requirements because his job was in a project to develop the requirements and specifications for a new generation of computers. NCR, having mechanical machines, could see the writing on the wall in terms of forthcoming computers. His autobiography answers one question I had always wondered about - why locate ISDOS in the Department of Industrial and Operations Engineering at UoM, and not the Computing Department? His response is that he found it extremely valuable to see the problems he was grappling with as engineering problems rather than computing problems. His whole knowledge transfer approach was driven by his belief that merely proposing new ways to do system development would not attract much attention, but offering to provide software to try out while the research was being done, was far more attractive - and of course in the event he was right. Everyone benefitted - this was not just knowledge transfer from university to industry - it was bi-directional transfer - a whole extended learning community where everyone obtained value.
Closing Years - PRISE and University of Michigan (1991-2003)
In 1989 Dan made an important contribution to the conference 'Taking CASE into the 1990's in Geneva, Switzerland, where the Symposium speakers were Barry Boehm, Gary Brooks, Geoffrey Darnton, Roger Pressman, David Stone, and Daniel Teichroew (Teichroew 1989).
In the early 1990s, Dan and I started work on a book based on the 1989 symposium in Geneva. The last complete draft is dated 1993 (Teichroew and Darnton 1993). That was a substantial book. An earlier working title of that book was even assigned an ISBN (0-13-119074-1 - still showing up in searches!) when it was thought the manuscript was nearing completion. For a variety of reasons, that book remains unpublished.
By the time Meta Systems Ltd was taken over by LBMS the ISDOS sponsors had virtually disappeared apart from some who continued to be users of PSL/PSA. Presumably that was a very unfortunate consequence of spinning out the project into a private company; why would the sponsors continue to donate their tool testing and ideas to a private company rather than a university research project?
At least one paper has been re-published posthumously, and doubtless others will follow (Oh et al. 2007).
Dan died peacefully in Ann Arbor, Michigan on 8 July 2003 after a short illness. His UoM department published a memorial the following day (UoM-IoE 2003).
The future? - what future? - PSL/PSA has not been available to new users since about 1992. In that time, many more approaches to system development have come and gone. Perhaps the most extensively used alternative is UML. However, that has not solved the fundamental problems in the past 20 years, which were in fact solved by PSL/PSA.
A favourite topic for Dan in seminars was the problem of 'the software crisis'. He refers to a paper by Knuth specifically dealing with software crisis in the 1960s. At the Geneva conference in 1989 he was able to comment "so we have had the software crisis now for many years, 25 years in fact". Now he would be able to change that to 45 years. Of course, his simple point was that in terms of building systems, we are not faced with a crisis, we are faced with a chronic. If I turn to the first Phase 1 report for the ISDOS project and the motivation for the project (Teichroew and Sibley 1969) we can find the comments that currently (remember we're talking about 1969) there are the major drawbacks in the then approaches to system development:
- systems do not work correctly;
- they are very costly to build
- the requirements are not stated explicitly;
- the construction starts before the design is completed;
- few alternatives are examined in the design phase;
- the process takes too long;
- the programs are the only documentation;
- the procedures become inefficient as changes occur;
- the accommodation of changes is expensive.
This all sounds as familiar today as it did more than 40 years ago! We really are dealing with a chronic. Losing PSL/PSA was a great loss to the system development world.
ACM - Association for Computing Machinery
CCITT - Comité Consultatif International Téléphonique et Télégraphique
DEC - Digital Equipment Corporation
IBM - International Business Machines
IEEE - Institute of Electrical and Electronics Engineers
ISDOS - Information System Design and Optimization System
ISLDS - Information System Language Definition System
ISETT - Information Systems Engineering Tools and Technology
LBMS - Learmonth and Burchett Management Systems
LDM - Language Definition Manager
NBS - National Bureau of Standards
PRISE - Program for Research in Information Systems Engineering
PSA - Problem Statement Analyzer
PSL - Problem Statement Language
SEM - System Encyclopedia Manager
SODA - Systems Optimization and Design Algorithm
SSADM - Structured Systems Analysis and Design Method
TRW - Thompson Ramo Wooldridge Inc.
UML - Unified Modeling Language
UoM - University of Michigan
Ali, I., Chan, Y., Kuick, R., Teichroew, D., and Hanash, S. M. (1991). "Implementation of a two-dimensional electrophoresis-related laboratory information processing system: database aspects." Electrophoresis, 12(10), 747-61.
Ashenhurst, R. L. (1972). "Curriculum Recommendations for Graduate Professional Programs in Information Systems." Communications of the ACM, 15(5), 363-398.
Block, C., Kang, K. C., Chikofsky, E., and Teichroew, D. (1983). Usage of the System Encyclopedia Manager (SEM) System with the CCITT Functional Specification and Description Language (CCITT/SDL). ISDOS Project Report MO399, Ann Arbor, MI.
Boydstun, L. E., Teichroew, D., Spewak, S., Yamamoto, Y., and Starner, G. (1980). "Computer Aided Modeling of Information Systems"IEEE Computer Society Conference. City: IEEE Computer Society, pp. 37-41.
Bull, W. E., Africa, C., and Teichroew, D. (1955). "Some Problems of the 'Word'", in W. W. Locke and A. D. Booth, (eds.), Machine Translation of Languages. New York and London: The Technology Press of The Massachusetts Institute of Technology and John Wiley & Sons, Inc., Chapman & Hall Ltd., pp. pp 86-103.
Canning, R. G. (1956). Electronic Data Processing for Business and Industry, New York: John Wiley & Sons.
Carnap, R. (1928). Der Logische Aufbau der Welt, Berlin-Schlachtensee: Weltkreis Verlag.
Carnap, R. (1967). The Logical Structure of the World and Pseudoproblems in Philosophy, London: Routledge & Kegan Paul.
Chaar, J. K., Teichroew, D., and Volz, R. A. (1993a). "Developing Manufacturing Control Software: A Survey and Critique." International Journal of Flexible Manufacturing Systems, 5(1), 53-88.
Chaar, J. K., Teichroew, D., and Volz, R. A. (1993b). "Real-time software methodologies: Are they suitable for developing Manufacturing control software?" International Journal of Flexible Manufacturing Systems, 5(2), 95-128.
Couger, J. D., Colter, M. A., and Knapp, R. W. (1982). Advanced System Development/Feasibility Techniques, New York: John Wiley & Sons.
Couger, J. D., and Knapp, R. W. (eds.), (1974)."System Analysis Techniques". New York: John Wiley & Sons, pp. 509 + xii
Darnton, G. (1987). "Automating the Generation of Functional Specifications"CASE Studies Conference, Ann Arbor, May 1987. City: PRISE Project, University of Michigan: Ann Arbor, MI.
Darnton, G. (2002). "Modelling Requirements and Architecting Large-Scale On-Line Competence-Based Learning Systems", in P. Kommers, V. Petrushin, Kinshuk, and I. Galeev, (eds.), Proceedings of the 2nd IEEE International Conference on Advanced Learning Technologies (ICALT 2002). Kazan, Tatarstan, Russia: IEEE.
Darnton, G., and Maddison, R. (1985). "The Use of the ISDOS PSL/PSA System with the Digital VAX Information Architecture Product Set"ISDOS/PRISE/IDA European Meeting. City: PRISE Project, University of Michigan: Paris.
Dixon, W. J., and Teichroew, D. (1953). Some Sampling Results on the Power of Nonparametric Tests against Normal Alternatives - working paper. National Bureau of Standards, Washington, DC.
Drennan, L. T., and Beck, M. (2001). "Teaching quality performance indicators - key influences on the UK universities' scores." Quality Assurance in Education, 9(2), pp 92-102.
Foster, F. G., and Teichroew, D. (1955). "A Sampling Experiment on the Powers of the Records Tests for Trend in a Time Series." Journal of the Royal Statistical Society Series B, 17, 6.
Hanash, S. M., Strahler, J. R., Chan, Y., Kuick, R., Teichroew, D., Neel, J. V., Hailat, N., Keim, D. R., Gratiot-Deans, J., Ungar, D., and et al. (1993). "Data base analysis of protein expression patterns during T-cell ontogeny and activation." Proc Natl Acad Sci U S A, 90(8), 3314-8.
Hanash, S. M., and Teichroew, D. (1998). "Mining the human proteome: experience with the human lymphoid protein database." Electrophoresis, 19(11), 2004-9.
Hirsch, R. S., Teichroew, D., and Paige, J. C. (1960). Experimental Evaluation and Statistical Analysis of Human Performance in a Constrained Handwriting Method of Data Recording. IBM Advanced Systems Development Division Laboratory, San Jose, CA.
Jenkins, A. (2004). A Guide to the Research Evidence on Teaching-Research Relations. Oxford: Higher Education Academy.
Kang, K. C., Darnton, G., and Teichroew, D. (1984). The Use of the ISDOS PSL/PSA System with the Digital VAX Information Architecture System (VAX-IA) Product Set. Ann Arbor, MI.
McCluskey, W. A. (1969). Toward the Automatic Design of Data Organization for Large-Scale Information Processing Systems, PhD Thesis, Case Western Reserve University, Cleveland, OH.
McGee, W. C. (1963). "The Formulation of Data Processing Problems for Computers", in F. L. Alt and M. Rubinoff, (eds.), Advances in Computers vol IV. Academic Press, pp. 1-52.
Merten, A., and Teichroew, D. (1972). "The impact of problem statement languages on evaluating and improving software performance"AFIPS 1972 Fall Computer Joint Conference. City: AFIPS Press: Anaheim, CA, pp. 849-857.
Nunamaker, J. F. (1969). On the Design and Optimization of Information Processing Systems, PhD Thesis, Case Western Reserve University, Cleveland, OH.
Oh, J. M., Hanash, S. M., and Teichroew, D. (1999). "Mining protein data from two-dimensional gels: tools for systematic post-planned analyses." Electrophoresis, 20(4-5), 766-74.
Oh, J. M. C., Hanash, S. M., and Teichroew, D. (2007). Mining protein data from two-dimensional gels: Tools for systematic post-planned analyses: Wiley-VCH Verlag GmbH.
Prince, M. J., Felder, R. M., and Brent, R. (2007). "Does Faculty Research Improve Undergraduate Teaching? An Analysis of Existing and Potential Synergies." Journal of Engineering Education, 96(4), pp 283-294.
Robichek, A. A., and Teichroew, D. (1961). "An Investment Decision in a Life insurance Company." Management Technology, 1, pp24-35.
Robichek, A. A., Teichroew, D., and Jones, M. (1965). "Optimal Short-Term Financing Decision." Management Science, 12(1), pp 1-36.
Robichek, A. A., Teichroew, D., and Jones, M. (1967). "Optimal Short-Term Financing Decision", in P. Hunt and V. Andrews, (eds.), Financial Management: Cases and Readings. Holmwood, IL: Richard D. Irwin.
Sammet, J. E. (1969). Programming Languages: History and Fundamentals, Englewood Cliffs, NJ: Prentice-Hall, Inc.
Sayani, H. H. (1990). "PSL/PSA at the Age of Fifteen", in R. H. Thayer and M. Dorfman, (eds.), System and Software Requirements Engineering. Los Alamitos, CA: IEEE Computer Society Press, pp. 403-417.
Spewak, S. H. (1981). Analysis of Dynamics of the Logical Design of Information Systems, PhD Thesis, University of Michigan, Ann Arbor, MI.
Steiger, W., and Teichroew, D. (1968). A Problem Statement Language Preliminary User's Manual". Case Western Reserve University, Cleveland, Ohio.
Teichroew, D. (1944). "Royal Canadian Airforce Flying Log Book for Aircrew other than Pilot". City.
Teichroew, D. (1951). "Empirical Sampling Distributions II"Meeting of the Institute of Statistics, University of North Carolina. City: National Bureau of Standards Report 1621: Boston, MA, pp. 20.
Teichroew, D. (1952). "The Use of Continued Fractions in High-Speed Computing." Mathematical Tables and Other Aids to Computation, VI(39), pp 127-133.
Teichroew, D. (1953a). Distribution Sampling with High Speed Computers, North Carolina State College, subsequently North Carolina State University.
Teichroew, D. (1953b). Experiment Designs - working paper. National Bureau of Standards, Washington, DC.
Teichroew, D. (1954a). Statistical Analysis of Experimental Parachute Test Data. National Bureau of Standards, Washington, DC.
Teichroew, D. (1954b). A Table of Campbell's Polynomials for the Computation of Probability Points of the Incomplete Gamma Distribution - working paper. National Bureau of Standards, Washington, DC.
Teichroew, D. (1955a). "Empirical Power Functions for Nonparametric Two-Sample Tests for Small Samples." Annals of Mathematical Statistics, 26, 5.
Teichroew, D. (1955b). "Numerical Analysis Research Unpublished Statistical Tables." Journal of the American Statistical Association, 50, 7.
Teichroew, D. (1955c). Probabilities Associated with Order Statistics in Samples From Two Normal Populations with Equal Variance. Army Chemical Center, Maryland.
Teichroew, D. (1956). "Tables of Expected Values of Order Statistics and Products of Orders Statistics for Samples of Size Twenty and Less from the Normal Distribution." Annals of Mathematical Statistics, 27, pp 410-426.
Teichroew, D. (1957). "The Mixture of Normal Distributions with Different Variances." Annals of Mathematical Statistics, 28(2), 510-512.
Teichroew, D. (1958a). "Automation of Information Processing", in E. E. Department, (ed.), Proceedings of the EIA Conference on Automation Systems for Business and Industry. New York: Engineering Publishers, pp. 180 + ix.
Teichroew, D. (1958b). "Decision Factors Influencing Adoption of Automated Business Systems"Proceedings of the Second EIA Conference on Automation Systems. City: Engineering Publishers, pp. pp 91-98.
Teichroew, D. (1959). Computers in Business. Stanford Research Institute.
Teichroew, D. (1960). "The Impact of Computers on Management." Stanford Graduate School Bulletin, 29(2), 5.
Teichroew, D. (1962). "Tables of Lower Moments of Order Statistics for Samples from the Normal Distribution", in A. E. Sarhan and B. G. Greenberg, (eds.), Contributions to Order Statistics. New York: John Wiley & Sons, pp. 190-199.
Teichroew, D. (1963). "Data Display in Business Information Systems." Stanford Graduate School of Business Bulletin, 32(3).
Teichroew, D. (1964a). An Introduction To Management Science Deterministic Models, New York: John Wiley & Sons.
Teichroew, D. (1964b). "Papers presented at the third annual symposium of the San Francisco Bay Area chapter of ACM." Commun. ACM, 7(7), 425.
Teichroew, D. (1965a). "The Application of Mathematics and Statistics in Business"Data Processing Management Association 1965 Fall Conference and Business exposition. City, pp. pp 410-414.
Teichroew, D. (1965b). "A History of Distribution Sampling Prior to the Era of the Computer and Its Relevance to Simulation." Journal of the American Statistical Association, 60, pp 27-49.
Teichroew, D. (1965c). "Information Systems in Information Display." Information Display, 2(6), pp 33-39.
Teichroew, D. (1966). "Case Institute Program on Organizational Science"Northwest Universities Business Administration Conference 1966. City, pp. pp 55-77.
Teichroew, D. (1967a). ISDOS - A Research Project to Develop Methodology for the Automatic Design and Construction of Information Processing Systems. Case Western Reserve University, Cleveland, Ohio.
Teichroew, D. (1967b). "Major Decisions Regarding Computers in Colleges and Universities." Association for Educational Data Systems Journal, 1(1), pp 25-45.
Teichroew, D. (1968). "ACM's Role in Business Data Processing." ACM Special Interest Group in Business Data Processing, pp 1-6.
Teichroew, D. (1969a). "ISDOS Information Systems Design and Optimization System"Diebold Research Program, Twenty-First Regular Meeting. City: Diebold Research Program: New York City.
Teichroew, D. (1969b). "Methodology for the Design of Information Processing Systems"Fourth Australian Computer Conference. City.
Teichroew, D. (1969c). "Scientific Methods in Commerce and Industry"Fourth Australian Computer Conference. City.
Teichroew, D. (1970a). "Problem Statement Languages in MIS", in E. Grochla and N. Szyperski, (eds.), Management-Informations-systeme. Wiesbaden: Betriebswirtschaftlicher Verlag Dr. Th. Gabler, pp. 251-282.
Teichroew, D. (1970b). "What Should MIS Do?"TIMS. City: Washington, DC.
Teichroew, D. (1971a). "An approach to research in file organization"Proceedings of the 1971 international ACM SIGIR conference on Information storage and retrieval. City: ACM: College Park, Maryland, pp. 147-154.
Teichroew, D. (1971b). "Education related to the use of computers in organizations." Commun. ACM, 14(9), 573-588.
Teichroew, D. (1972a). "Computer-Aided Documentation of User Requirements", in F. J. Gruenberger, (ed.), Information Systems for Management. Englewood Cliffs: Prentice-Hall Inc, pp. pp 97-112.
Teichroew, D. (1972b). "A survey of languages for stating requirements for computer-based information systems"Proceedings of the December 5-7, 1972, fall joint computer conference, part II. City: ACM: Anaheim, California, pp. 1203-1224.
Teichroew, D. (1974a). "Improvements in the System Life Cycle"Information Processing 74. City: North-Holland Publishing Company, pp. pp 972-978.
Teichroew, D. (1974b). "Problem Statement Analysis: Requirements for the Problem Statement Analyzer (PSA)", in J. D. Couger and R. W. Knapp, (eds.), System Analysis Techniques. New York: John Wiley & Sons, pp. pp 336-358.
Teichroew, D. (1974c). "Problem Statement Languages in MIS", in J. D. Couger and R. W. Knapp, (eds.), System Analysis Techniques. New York: John Wiley & Sons, pp. pp 310-327.
Teichroew, D. (1982). "The Development of Software Support Environments", Proceedings of Canadian Information Processing Society Session 82. Saskatoon, Saskatchewan, Canada: Canadian Information Processing Society, pp. 200-210.
Teichroew, D. (1985). "Computer Aided Tools to Support Methodologies for Enterprise-Wide Information Management"Second International Workshop on Enterprise-Wide Information Management. City: Washington University, St. Louis.
Teichroew, D. (1989). "System Engineering and CASE in the 1990s", G. Darnton, (ed.) Taking CASE into the 1990s. City: Information Architecture Program Office, Digital Equipment Corporation International (Europe): Geneva, Switzerland, pp. 4-1 to 4-87.
Teichroew, D. (1992). "J.-D. Warnier Prize Winner Autobiographical Sketch". City: J.-D. Warnier Prize Committee.
Teichroew, D. (1993). "If "Software Process" Is the Answer, What Was the Question?"Fifth Annual Montertey Software Conference. City: Digital Consulting Inc.: Monterey, CA.
Teichroew, D., and Chikofsky, E. (1980). "An Introduction to The Problem Statement Language (PSL) and The Problem Analyzer (PSA)"Share 55. City: ISDOS Project: Atlanta, GA.
Teichroew, D., and Darnton, G. (1993). "System and Software Development into the 21st Century - manuscript draft". City: Fareham, England.
Teichroew, D., and Dávid, G. (1985). System description methodologies : proceedings of the IFIP TC2 Conference on System Description Methodologies, Kecskemét, Hungary, 23-27 May, 1983, Amsterdam ; New York New York, N.Y.: North-Holland.
Teichroew, D., and Erdman, R. L. (1963). Measurement of Print Chacteristics. IBM Advanced Systems Development Division, San Jose, CA.
Teichroew, D., and Gackowski, Z. (1977). "Structured systems design"Annual Conference for the Association for Systems Management. City, pp. 45-58.
Teichroew, D., Germano, F., and Silva, L. (1981). "Applications of The Entity-Relationship Approach", in P. P. Chen, (ed.), Entity-Relationship Approach to Information Modeling and Analysis. Saugus, CA: ER Institute, pp. 1-17.
Teichroew, D., and Hershey, E. A. (1977). "PSL/PSA: A Computer-Aided Technique for Structured Documentation and Analysis of Information Processing Systems." IEEE Transactions on Software Engineering, 3(1), pp 6-15.
Teichroew, D., and Hershey, E. A. (1982). "PSL/PSA: A Computer-Aided Technique for Structured Documentation and Analysis of Information Processing Systems", in J. D. Couger, M. A. Colter, and R. W. Knapp, (eds.), Advanced System Development/Feasibility Techniques. New York: John Wiley & Sons.
Teichroew, D., Hershey, E. A., and Yamamoto, Y. (1982a). "The PSL/PSA Approach to Computer-Aided Analysis and Documentation", in J. D. Couger, M. A. Colter, and R. W. Knapp, (eds.), Advanced System Development/Feasibility Techniques. New York: John Wiley & Sons, pp. 330-346.
Teichroew, D., Hershey, E. A., and Yamamoto, Y. (1982b). "The PSL/PSA Approach to Computer-Aided Analysis and Documentation", in J. D. Couger, M. A. Colter, and R. W. Knapp, (eds.), Advanced System Development/Feasibility Techniques. New York: John Wiley & Sons.
Teichroew, D., and Kang, K. C. (1984). "A Software Environment to Support the CCITT Development Methodology"Third CHILL Conference. City: Cambridge University, England, pp. 125-132.
Teichroew, D., Kang, K. C., and Macasovic, P. (1980a). "A Laboratory for Software Engineering Education and Research", Proceeedings of II International Symposium on the Computer at the University. Zagreb, Yugoslavia: Zagreb University Computing Center, pp. I-38 to I-49.
Teichroew, D., Lesso, W., Rice, K., and Wright, G. (1967a). "Optimizing Models of After-Tax Earnings Incorporating Depletion Allowances." Journal of Financial and Quantitative Analysis, 2, pp 265-297.
Teichroew, D., Lubin, J., and Truitt, T. (1967b). "Discussion of computer simulation techniques and comparison of languages." Simulation, 9(4), 181-190.
Teichroew, D., and Lubin, J. F. (1966). "Computer Simulation - Discussion of the Technique and Comparison of Languages." Communications of the ACM, 9(10), pp 723-741.
Teichroew, D., Macasovic, P., Hershey, E. A., and Yamamoto, Y. (1980b). "Application of the Entity-Relationship Approach to Information Processing Systems Modelling", in P. P. Chen, (ed.), Entity-Relationship Approach to Systems Analysis and Design. Amsterdam: North-Holland Publishing Co., pp. pp 15-30.
Teichroew, D., and Miller, B. (1964). A Generalized Dynamic Programming Routine. Stanford Graduate School of Business.
Teichroew, D., Nielson, N., and Lubin, J. F. (1964a). An Empirical Evaluation of Technical Stock Market Investment Rules. Stanford Graduate School of Business, Stanford.
Teichroew, D., Perry, N. C., and Caffrey, J. (1956). "An Empirical Study of an Estimate of Skewness for the Pearson Type III Distribution." American Statistician, 10, 2.
Teichroew, D., Rataj, W. J., and Hershey, E. A. (1975). "An Introduction to Computer Aided Documentation of User Requirements for Computer Based Information Processing Systems", in Grochla and Szyperski, (eds.), Information Systems and Organization Structure. Berlin: Walter De Gruyter.
Teichroew, D., Robichek, A. A., and Montalbano, M. (1965a). "An Analysis of Criteria for Investment and Financing Decisions Under Certainty." Management Science, 12(3), pp 151-179.
Teichroew, D., Robichek, A. A., and Montalbano, M. (1965b). "Mathematical Analysis of Rates of Return Under Certainty." Management Science, 11(3).
Teichroew, D., Robichek, A. A., and Montalbano, M. (1968). "An Analysis of Criteria for Investment and Financing Decisions Under Certainty", in G. P. E. Clarkson, (ed.), Managerial Economics: Selected Readings. Harmondsworth: Penguin Books.
Teichroew, D., and Sayani, H. (1971). "Automation of System Building"Datamation. City, pp. pp 25-20.
Teichroew, D., and Sayani, H. (1974). "Automation of System Building", in J. D. Couger and R. W. Knapp, (eds.), System Analysis Techniques. New York: John Wiley & Sons, pp. pp 379-389.
Teichroew, D., and Sayani, H. (1980). "Computer-aided requirements engineering"Proceedings of the ACM 1980 annual conference. City: ACM, pp. 369-381.
Teichroew, D., and Sibley, E. H. (1969). ISDOS - A Research Project to Develop Methodology for the Automatic Design and Construction of Information Processing Systems and Phase I Report University of Michigan, Ann Arbor.
Teichroew, D., and Sibley, E. H. (1970). "Automatic Systems Building Seminar"Computers in the 70s Seminars. City: Urwick Dynamics Ltd.: Manchester and London.
Teichroew, D., and Sitgreaves, R. (1957). Probability and Statistics in Item Analysis and Classification Problems., USAF School of Aviation Medicine.
Teichroew, D., and Sitgreaves, R. (1961). "Computation of an empirical sampling distribution for the classification statistic W.", in H. Solomon, (ed.), Studies in Item Analysis and Prediction. Stanford: Stanford University Press.
Teichroew, D., Smith, J. M., and Snell, E. (1963). Stanford Mechanized Accounting Cases. Stanford University, Palo Alto, CA.
Teichroew, D., Smith, J. M., and Snell, E. (1964b). Computerized Practice Set: SMAC Corporation, New York: McGraw-Hill Book Co.
Teichroew, D., Spewak, S., Hershey, E. A., Yamamoto, Y., and Starner, G. (1979). "Computer Aided Modeling of Information Systems"IEEE Compsac Conference. City: Chicago.
Teichroew, D., and Winters, E. W. (1976). "Recent Developments in System Analysis and Design." Atlanta Economic Review, pp 39-46.
Tremblay, E. J. P. (1969). A Problem Statement Language Definition and its Syntactic Analysis, PhD Thesis, Case Western Reserve University, Cleveland, OH.
Umar, A., and Teichroew, D. (1990). "Pragmatic issues in conversions of database applications." Information & Management, 19, 149-166.
University, N. C. S. (1953). "1953 PhD and Master's Recipients ". City: North Carolina State University: Raleigh, NC, pp. 1953 Alumni - PhD and Master's.
UoM-IoE. (2003). "Memorial to Daniel Teichroew". City: University of Michigan Department of Industrial and Operations Engineering: Available at: http://ioe.engin.umich.edu/newsandevents/page.php?id=95&archive=yes.
Wegner, P., Dennis, J., Hammer, M., and Teichroew, D. (eds.), (1979)."Research directions in software technology"MIT Press series in computer science 2. Cambridge, Mass.: MIT Press: Cambridge, Mass., pp. xiii, 869 p.
Welke, R. J. (1986). "Open Letter from Meta Systems"Meta Systems Developments. City: Meta Systems Ltd.: Ann Arbor, MI, pp. 1.
Winters, E. W. (1981). On the Analysis and Synthesis of Human and Computer Subsystems for Information Processing Systems, PhD Thesis, Union of Experimental Colleges and Universities, Cincinnati, OH.
Yamamoto, Y. (1981). An Approach to the Generation of Software Life Cycle Support Systems, PhD Thesis, University of Michigan, Ann Arbor.
Young, J. W., and Kent, H. K. (1958). "Abstract Formulation of Data Processing Problems." Journal of Industrial Engineering, IX(6), pp 471-479.