A lack of historical awareness invites repeated astonishment: developments whose histories are forgotten seem to arise out of nowhere. The "computer age" as a social movement, with its relentless marketing-oriented focus on the future, is particularly prone to this erosion of collective memory. Yet it is a product of its history, even as it tends to forget the foundations on which it rests.

This collection of information aims to outline the wider context of the interdependent influences that have shaped Logo and its legacy. It is not intended as a comprehensive history of every implementation or project, but as a selective account of the people, ideas, technologies, and educational ambitions that intersect with the development of Logo.

For a brief overview of the broader history of artificial-intelligence research, with further examples of early projects and their context, see MIT CSAIL's collection of AI films.

Prehistory

1930s

1931 Kurt Gödel, mathematician and logician, publishes his paper on formally undecidable propositions in Principia Mathematica and related systems, establishing his first incompleteness theorem: sufficiently powerful, consistent formal systems cannot be both complete and consistent in the relevant sense.

1935 Alonzo Church, mathematician, develops and presents his lambda calculus, work that also engages with questions raised by Gödel's incompleteness theorems (Sieg 1996, pp. 155–156).

1936 Church's former student Stephen Cole Kleene, mathematician, publishes General Recursive Functions of Natural Numbers, a contribution to the development of recursion theory in the context of questions concerning formal calculability.

Alan M. Turing publishes On Computable Numbers, with an Application to the Entscheidungsproblem. He introduces an abstract model of computation, subsequently known as the Turing machine. Turing's model, Church's lambda calculus, and Kleene's recursive-function approach provide equivalent accounts of computability. Together, these approaches become foundational to theoretical computer science and the modern understanding of computation.

1937 Jean Piaget, biologist and psychologist, publishes La construction du réel chez l'enfant (The Construction of Reality in the Child), a study of how children construct their understanding of the world.

1940s

1942 The Macy Foundation's first Cerebral Inhibition Meeting, an interdisciplinary conference concerned with neural systems, takes place. Among its invited participants is Warren S. McCulloch, psychiatrist and neurophysiologist (Edwards 1996, p. 181; Heims 1993; cited in N.A. 2023).

1943 McCulloch and Walter Pitts, a mathematician and logician, publish A Logical Calculus of the Ideas Immanent in Nervous Activity. They present a mathematical model of neural activity in terms of propositional logic. Their work becomes one of the foundations of research into artificial neural networks and machine learning. The model also has a formal relationship to the Turing machine (Edwards 1996, p. 188).

Pitts begins his studies at the Massachusetts Institute of Technology (MIT) under Norbert Wiener, mathematician.

Together with Arturo Rosenblueth Stearns, physiologist and electrical engineer, Wiener publishes Behavior, Purpose and Teleology. The paper examines how goal-directed behaviour can be understood through feedback and circular causal relations.

1944/45 John von Neumann, mathematician, and Wiener organise further meetings involving mathematicians, engineers, neuroscientists and social scientists. These meetings provide a forum for discussing common features of natural nervous systems and artificial computing machines. Their aim is to establish a foundation for a new scientific discipline capable of describing natural and technical systems within a common framework (Edwards 1996, p. 189).

1946 McCulloch organises a series of conferences on these topics under the auspices of the Macy Foundation. Ten conferences in this series, subsequently known as the Macy Conferences, take place between 1946 and 1953 (Edwards 1996, p. 189).

Heinz von Foerster is invited to participate in a conference, where he gives a much-discussed presentation on a quantum-physical theory of memory; he is subsequently appointed editor of the conference proceedings (Edwards 1996, p. 189). In 1967, another participant in the conferences, Margaret Mead, calls for cybernetics to be applied reflexively to the practice of cybernetics itself. Her address, published in 1968 as Cybernetics of Cybernetics becomes an important reference point for the development of second-order cybernetics (von Foerster 2003, p. 302; Mead 1968. In: von Foerster et al. (eds.), p. 1-11). This epistemological orientation has affinities with radical constructivism.

1947 Norbert Wiener publishes his open letter, “A Scientist Rebels”, refusing to provide a researcher working on guided missiles with a copy of his wartime research paper. Acknowledging that the paper could still be obtained through other channels, Wiener describes his refusal as a pro forma protest against the military use of scientific knowledge. He also declares that he will not publish future work that might cause harm in the hands of irresponsible militarists.

1948 Claude Shannon, mathematician and engineer at AT&T Bell Laboratories, publishes A Mathematical Theory of Communication, laying the foundations of information theory.

Wiener publishes Cybernetics: Or Control and Communication in the Animal and the Machine, synthesising aspects of the ongoing interdisciplinary discourse. The book becomes a major theoretical foundation for subsequent developments in cybernetics and artificial intelligence. Wiener derives the term cybernetics from the ancient Greek κυβερνήτης (kybernētēs), meaning steersman, and presents cybernetics as an interdisciplinary study of control and communication in animals and machines. His work also anticipates problems associated with these developments—problems that, more than seventy years later, are still sometimes presented as novel and unprecedented.

W. Ross Ashby, psychiatrist and cybernetician, presents the Homeostat, an adaptive mechanical apparatus of his own design. The device acts as a self-regulating system and applies a principle associated with physiological homeostasis (Wiener 2021, p. 190; The Science Team 2016).

1948–1949 William Grey Walter, neurophysiologist, develops two cybernetic robots, known as tortoises, which navigate their environment autonomously (Marsh 2020).

Around the same time, Wiener and Henry Earl Singleton also construct a robot named Palomilla (Spanish for “moth”). Its control system enables it to respond autonomously to light (Hoggett 2009).

1949 Von Neumann implements his conception of an electronic general-purpose computer at the Institute for Advanced Studies; his Von-Neumann machine is a foundation of computer technology until today (Burks 1966. In: von Neumann, pp. 8-10).

1950s

1950 Turing publishes his paper Computer Machinery and Intelligence in which he develops the thought experiment of the Imitation Game, later commonly known today as the Turing test.

Wiener publishes The Human Use of Human Beings as a critical perspective on the societal effects of cybernetics.

1951 Shannon presents a mechanic, cybernetic mouse which autonomously moves through a maze (Hoggett 2009).

1952 McCulloch starts teaching at the MIT where Wiener arguably fell out with Pitts and him irreconcilable (Gefter 2015).

1953 Von Neumann develops a theory of self-reproducing automata (Burks 1966. p. xv).

1954 John W. Backus and colleagues start developing an expressive—high-level—programming language, which can be used—plattform independently: Fortran (Backus 1998)

Around the same time, the soviet mathematician Ekaterina Yushchenko and colleagues start developing the likewise plattform independent high-level programming language Адресный язык программирования (Address Programming Language) (Mullin/Hains 2026; Glushkov/Yushchenko 1966/1962; Fels 1962).

1955 John McCarthy, mathematician, starts organising the 1956 Dartmouth Summer Research Project with the mathematician Marvin Minsky, who obtained his doctoral degree with a thesis on neural networks a year before, as well as with Shannon, and with Nathaniel Rochester, engineer, who substantially contributed to the design of the first computer of IBM, developed for research purposes. In the proposal, McCarthy uses the term artificial intelligence for the first time, meant to attract researcher's in the field and as distinction from cybernetics associated with Wiener (Anderson 2024, p. 4; N.A. 2019, p. 385; Moore 2006, p. 87; McCarthy 2006, p.12).

1956 Ashby publishes An Introduction to Cybernetics, an accessible introduction to the emerging discipline.

Allen Newell, John Cliford Shaw and Herbert A. Simon, two mathematician and a political scientist, demonstrate their Logic Theorist at Dartmouth. It is a program they developed at the RAND Corporation, designed to prove theorems in propositional logic. Its implementation draws on IPL (Information Processing Language), an assembly-language-based programming language developed for symbolic information processing (Newell/Tonge 1960, p. 2).

J.C.R. Licklider, a psychologist with interdisciplinary interests spanning psychoacoustics, mathematics, and computing, joins Bold, Beranek and Newman Inc. (BBN), an engineering consultancy with close ties to MIT. His work increasingly concerns the interaction between people and computer systems, including the use of computers to support learning, problem-solving, and decision-making

McCarthy moves from Dartmouth College to MIT (Rajamaran 2014, p. 203).

1957 Frank Rosenblatt publishes his report on Project PARA, titled: The Perceptron: A Perceiving and Recognizing Automaton. Drawing on research in physiological psychology, neural-network models, and statistical approaches to learning and recognition, Rosenblatt proposes a system intended to model aspects of perception and learning. The perceptron becomes an important precursor of modern neural networks, although its capabilities and limitations will remain subjects of substantial debate.

FORTRAN I is presented at the Western Joint Computer Conference in Los Angeles (Backus 1978, p. 40).

European and American mathematician and engineers begin working towards an international standard for a high-level programming language, building on and seeking to improve upon existing languages such as FORTRAN. This collaboration leads to ALGOL (Perlis 1978).

On 4 October, the Soviet Union launches Sputnik 1, the first artificial satellite to orbit Earth. The launch intensifies concerns in the United States and other NATO countries about scientific and technological competition known as the Sputnik crisis. In the United States, it contributes to substantial changes in defence research and science and mathematics education, including new federal funding initiatives (Solomon et. al. 2020, p. 20).

1958 In response to the perceived strategic implications of the Sputnik launch, the US Department of Defense establishes the Advanced Research Projects Agency (ARPA) on 7 February to support research and development in areas of potential strategic importance (Dennis 2024).

Minsky joins MIT and, together with McCarthy, establishes the institutional foundations (Artificial Intelligence Group) of what later becomes the MIT Artificial Intelligence Laboratory (Winston 2016; Bernstein 1981; McCarthy 1979, p. 6).

Papert joins Piaget's Centre International d’Épistémologie Génétique in Geneva, Switzerland (Papert 1980, p. 208).

McCarthy develops the central ideas of LISP (LISt Processor), initially conceived as a formal language for manipulating symbolic expressions. To his surprise, Steve Russell, a programmer, realises that the eval procedure described in McCarthy's work can be implemented in machine code and writes an interpreter for it. Drawing on recursive function theory and Church's lambda calculus, and developed in the context of contemporary programming languages such as FORTRAN and IPL, LISP thus moves from a formal language design to an executable high-level programming language (Turner 2013, pp. 6–7; Graham 2002, p. 4; Sridharan 1982, p. 26; McCarthy 1960, p. 185).

The name ALGOL (ALGOrithmic Language) is adopted for the international programming-language project, initially known as IAL (International Algebraic Language) (Perlis 1978, p. 79).

Von Foerster establishs the Biological Computer Laboratory (BCL) at the University of Illinois, Urbana-Champaign. The laboratory becomes an interdisciplinary centre for research into cybernetics, self-organising systems, biological processes, and the relationships between living and artificial systems.

1959 Jerome Y. Lettvin, Humberto R. Maturana, McCulloch, and Pitts publish What the Frog's Eye Tells the Frog's Brain. Investigating the frog's optic nerve, they identify four functionally distinct types of nerve fibres that respond to different visual patterns, largely independently of overall illumination. Their findings challenge the view of vision as a passive registration of light intensity and suggest that sensory processing already involves the selective extraction of features from the visual field.

1960s

1960 In Man-Computer Symbiosis, Licklider outlines a vision of human–computer cooperation. He envisages computers handling routine operations and preparing information for decisions, while people retain responsibility for setting objectives and making judgements within the bounds they specify (Licklider 1960).

McCarthy publishes Recursive Functions of Symbolic Expressions and Their Computation by Machine, Part I, presenting LISP as a programming language and describing its computational foundations (McCarthy 1960). In the same year, the ALGOL 60 report establishes the definition of ALGOL 60 (Backus et al., 1960).

Historical note: ALGOL's legacy

Certain concepts introduced or formalised in ALGOL will also influence the further development of Lisp in the 1970s, notably through Scheme, a Lisp dialect whose design explicitly draws on ALGOL 60. The influence of ALGOL on Lisp is more than a matter of surface syntax: its block structure, for example, becomes an important element in Scheme's design (Abelson, Sussman & Sussman 1996; Pitman (ed.) 1996; Steele Jr. & Gabriel 1993, p. 249).

ALGOL's overall influence on the design of subsequent programming languages is profound. Its development was also connected with efforts to overcome limitations of existing languages, including FORTRAN, and to provide a more systematic and machine-independent means of expressing algorithms.

One possible way of distinguishing programming-language traditions, therefore, is to consider languages that resemble ALGOL more explicitly in their syntax and program structure, and those that incorporate concepts associated with ALGOL without displaying the same resemblance at the syntactic level. This is an analytical distinction rather than a strict division: language traditions overlap, and influences may be carried forward in ways that are not immediately apparent from a language's surface syntax.

ALGOL 60's influence, however, did not translate into equally widespread adoption as an implemented programming language. Niklaus Wirth, an engineer and computer scientist, collaborated with C. A. R. Hoare on ALGOL W (1966), a language that served as a basis for Wirth's later design of Pascal (1970). Named in tribute to Blaise Pascal, Pascal was designed in part for teaching structured programming. Wirth subsequently developed languages including Modula and Modula-2. Another language in the broader ALGOL tradition is Ada, named in tribute to Ada Lovelace, and designed by a team led by Jean Ichbiah. These languages may be regarded as prominent examples of the more explicitly ALGOL-like tradition.

1961 Norbert Wiener expands Cybernetics: Or Control and Communication in the Animal and the Machine with two supplementary chapters, “On Learning and Self-Reproducing Machines” and “Brain Waves and Self-Organizing Systems.” These chapters extend his discussion of learning and self-reproduction and revisit the implications of such developments for society, including their potential military uses (Wiener 2021, p. vii-viii).

Minsky publishes Steps Toward Artificial Intelligence, surveying the field's state of the art from his perspective (Minsky 1961).

Seymour Papert and Minsky begin a collaboration grounded in shared interests in intelligence, learning, and the work of Piaget. Their subsequent work at MIT brings questions about children's learning and cognitive development into sustained contact with research in artificial intelligence (Solomon et al. 2020, p. 8).

1962 Licklider, McCarthy and Minsky are closely involved in the development of time-sharing systems at BBN and MIT, an approach that allows multiple users to interact with a computer and share its computational resources. Later that year, McCarthy moves to Stanford University (Solomon et al. 2020, p. 10; Swets 2011. In: Walden/Nickerson (Eds.), pp. 28-29).

Licklider and Welden E. Clark, an engineer, publish On-Line Man-Computer Communication describing their work on interactive computer programs for learning and problem-solving. They present TUTOR 1, a vocabulary-learning program, as an example of computer-aided teaching, and discuss the further development of GRAPH EQUATION, a program intended to help learners explore relationships between algebraic equations and their graphical representations. They also present COPLANNER, an interactive program for hospital planning and design, as an example of computer-aided planning and design. Finally, PROGRAM GRAPH and MEMORY COURSE demonstrate ways of analysing program flow for educational purposes and supporting error recovery, including the stepwise execution of instructions—features that anticipate aspects of later debugging tools (Licklider/Clark 1962, pp. 115-120).

Terminological note:

Here, on-line and off-line refer to whether a user or process is connected to the main computer. The terms do not yet carry all the meanings they acquired later. In this historical context, direct interaction with the computer takes place on-line (Tanenbaum 2015, pp. 8-9).

Wallace Feurzeig, a mathematician, begins working under Licklider in BBN's Artificial Intelligence Department (Feurzeig 2011. In: Walden/Nickerson (Eds.), pp. 282-283).

Cynthia Solomon, trained as a historian, joins Minsky's AI Group as a secretary, with the aim of learning to program. She later moves to BBN, where she works on translating MENTOR, a medical expert system developed by Feurzeig, into LISP (Solomon et al. 2020, p. 9; Feurzeig 2011, p. 283).

Steve Russell and fellow members of MIT's Tech Model Railroad Club (TMRC) develop Spacewar! for the PDP-1. Conceived by Russell, Martin Graetz and Wayne Wiitanen, the game is expanded through contributions from other programmers. It becomes one of the earliest and most influential video games for a digital computer (Solomon et al. 2020, pp. 10-11; Monnens/Goldberg 2015, p. 124).

ARPA recruits Licklider to direct its Information Processing Techniques Office (IPTO), where he promotes the development of time-sharing systems and advances his vision of an Intergalactic Computer Network, a precursor in conception to the later ARPANET (Pelkey 2021; Swets 2011, pp. 29-30).

1963 The ARPA funded Project MAC ("Multiple Access Computer and Multiple Aided Cognition," originally "Mathematics and Computation", also "Man and Computer" though never "Macintosh") is initiated at MIT, originally targeting the ongoing development of efficient time-sharing systems and soon integrating Minsky's AI Group (Giacaglia 2022).

1963 Shaw, co-inventor of the Logic Theorist and a contributor to the development of IPL, begins developing JOSS (Johnniac Open-Shop System), an interactive, high-level programming language based on FORTRAN. Its purpose is to make the formulation and solution of numerical problems more convenient through direct interaction with a computer (Shaw 1964, p. 2).

John G. Kemeny and Thomas E. Kurtz, both mathematician at Dartmouth college, pursue the goal of making computers and programming accessible to university students across disciplines. They begin developing BASIC (Beginner's All-purpose Symbolic Instruction Code), a language designed for beginners and intended to support a broad range of applications, rather than numerical problem-solving alone. Its design draws on existing programming-language traditions, including FORTRAN and JOSS (Kemeny/Kurtz 1964). Following John McCarthy's advice, Kemeny and Kurtz also pursue the development of a time-sharing system, involving their students in the implementation of a multi-terminal computing environment. The resulting Dartmouth Time-Sharing System will make interactive computing accessible to students and faculty across the college.

Historical note: Early FORTRAN and BASIC—programming practices and their legacy

BASIC will become the first programming language for many programmers, particularly during the personal-computing era from the 1970s onwards. One of Microsoft's earliest products was a BASIC interpreter developed by Bill Gates and Paul Allen for the Altair 8800. Yet BASIC has also frequently been criticised for encouraging programming practices that make source code difficult to understand and maintain. Some of these problems are associated with the use of unrestricted jumps between numbered lines, a feature inherited from earlier programming traditions and also found in assembly languages (Backus 1998, p. 72; Dijkstra 1968, p. 147). (Backus 1998, p. 72; Dijkstra 1968, p. 147). Such jumps can facilitate the development of so-called spaghetti code.

The structured-programming movement sought to address these problems by promoting more disciplined control structures and program organisation. Later languages such as Pascal incorporated these principles into their design and teaching, although structured programming is not reducible to the use of any single language (Dijkstra 1970). Nor is the structured-programming approach equally appropriate to every programming language or style of expression: languages grounded in different computational and mathematical traditions may offer other means of expressing relationships and organising programs. See Harvey (1997, pp. 233-238) for a critique of treating structured programming as a universal paradigm.

Nevertheless, structured programming became highly influential, particularly as languages with an ALGOL-like syntax and explicit control structures became widespread in education and the software industry. These languages and their associated practices helped shape professional programming communities and the conventions through which programs were designed, taught, and maintained.

In this broader historical landscape, the development of Logo can be understood as taking place in a partly different tradition. As a tentative heuristic, one might distinguish a more pragmatic or engineering-oriented tradition, often associated with explicit control structures and procedural organisation, from a more mathematical tradition, in which the expression and transformation of symbolic relationships play a more visible role. This is not a strict classification: the traditions overlap, and neither pragmatic necessarily means undogmatic, nor mathematical necessarily means impractical.

1964 Papert joins Minsky's AI Group (Solomon 2020, p. 8).

BBN demonstrates an interactive, multi-user computer system that points towards the practical use of time-sharing in educational settings, including public schools (Feurzeig 1984).

Logo history in the narrower sense

1960s

1964 At BBN, TELCOMP is developed as an implementation of JOSS. Feurzeig subsequently extends TELCOMP with string-processing capabilities, producing STRINGCOMP as a language intended to support mathematics education (Feurzeig 2011, p. 290).

1965 Feurzeig establishes the Educational Technology Department (ETD) at BBN to investigate how interactive computing can support learning and teaching. The group's focus shifts from tutorial applications towards programming languages as educational environments. With government support, Feurzeig and his collaborators investigate the use of STRINGCOMP in schools at elementary and secondary levels.

Feurzeig regards a programming language specifically designed for educational purposes as necessary. He identifies the following requirements (Feurzeig 2011, p. 291):

1. Third-graders with very little preparation should be able to use it for simple tasks.
2. Its structure should embody mathematically important concepts with minimal interference from programming conventions.
3. It should permit the expression of mathematically rich non-numerical algorithms, as well as numerical one.

1966 Feurzeig and his BBN group, including Cynthia Solomon and Daniel Bobrow, begin developing LOGO in collaboration with Seymour Papert, who works with the group as a consultant. Bobrow, a former doctoral student in MIT's Artificial Intelligence Group, had completed his doctorate in 1964 and was heading BBN's AI Group.

The new language draws on Lisp, initially conceived by its developers as Baby Lisp or Lisp without parentheses. Feurzeig proposes the name LOGO, derived from the ancient Greek word λόγος (lógos), meaning thought, word or speech, reflecting the initial emphasis on manipulating words and sentences. The design combines this linguistic orientation with a conception of programming as a medium for exploring ideas.

Papert's contribution is particularly significant in shaping the direction of the project towards a programming language designed specifically for children. The transition from STRINGCOMP to LOGO is therefore not merely a technical modification: it also reflects a reconsideration of what programming might make possible in children's learning.
(Solomon et al. 2020, pp. 9, 33; Feurzeig 2011, p. 291)

Historical note: ELIZA and Joseph Weizenbaum

In 1966, Joseph Weizenbaum publishes ELIZA—A Computer Program for the Study of Natural Language Communication between Man and Machine in Communications of the ACM (Weizenbaum 1966). Developed at MIT, ELIZA processes textual input using keyword recognition and transformation rules. Its best-known script, DOCTOR, simulates a psychotherapist by reformulating a user's statements and responding with questions. Although the program does not understand its interlocutors in the human sense, some users attribute a remarkable degree of understanding to it.

Weizenbaum is disturbed by this response. His experience with ELIZA contributes to his growing criticism of unreflective faith in technological progress and artificial intelligence. In Computer Power and Human Reason: From Judgment to Calculation (1976), he argues that the question of what computers can be made to do must be distinguished from the question of what they ought to be entrusted with doing. His work thus offers an important critical perspective from within computer science itself, alongside the enthusiasm surrounding the development of artificial intelligence and educational computing.

See Harvey (1997, pp. 147 ff.) for a discussion of ELIZA in Logo, and Norvig (1992, Chapter 5) for another implementation in Common Lisp. Both accounts are very informative even for readers with little or no prior knowledge of the respective programming language.

1967-69 Working implementations of LOGO are developed and tested with schoolchildren. The initial work, supported by the U.S. Office of Naval Research, includes experiments with fifth- and sixth-graders. Further research, supported by the U.S. National Science Foundation, extends the work with children from primary school through seventh grade, informing the continued development of the language and its educational environment.
(Solomon et al. 2020, pp. 33-35; Feurzeig 2011, p. 291)

The project is evaluated by mathematics and science education researchers involved in the educational reform efforts of the 1960s: Andrew Gleason, Robert Davis, Max Beberman, and Robert Karplus (Solomon et al. 2020, p. 10). Although these researchers are associated with the New Math movement, LOGO was not designed within this framework. Papert later articulates criticisms of aspects of the movement (Papert 1980, p. 53).

The BBN and MIT participants publish their collaborative project's final report, Programming-Languages as a Conceptual Framework for Teaching Mathematics (1970).

1968 Douglas Engelbart and his team at the Stanford Research Institute (SRI) demonstrate the results of their research on augmenting human intellect. Their presentation showcases computer technologies that anticipate many features of interactive computing that became widespread during the 1980s and 1990s. The demonstration, widely known as “the mother of all demos”, includes a computer mouse, hypertext, shared document editing, and videoconferencing. The contrast between these achievements and the BBC's 1967 vision of the “home computer of tomorrow ” makes tangible the gap between experimental research and contemporary visions of personal computing.

1969 Papert and Feurzeig's former collaborator Solomon return to MIT, where both establish a Logo Group within Marvin Minsky's A.I. Group. The Logo Group and Feurzeig's ETD continue their Logo research independently (Solomon et al. 2020, p. 36; Feurzeig 2011, p. 294).

Minsky and Papert publish Perceptrons: An Introduction to Computational Geometry, in which they mathematically examine the capabilities and limitations of Rosenblatt's perceptron model (1957). They also discuss research on biological vision, including Lettvin et al.'s study of the frog's visual system (1959), situating it within the wider history of cybernetics and pattern recognition.

At MIT, a tethered drawing robot is invented.

Late 1960s-early 1970s At Xerox Palo Alto Research Center (Xerox PARC), Alan Kay, software engineer and musician, Dan Ingalls, software engineer, Adele Goldberg, computer scientist with research focus on computer-aided learning, and further collaborators begin developing SMALLTALK. Their work draws on several programming language traditions, including LISP and LOGO, as well as SIMULA. SMALLTALK is conceived not only as a programming language but also as part of a broader vision of interactive personal computing and educational environments (Kay 2019, p. 2; Kay 1970, p. 4). Engelbart's work at SRI provides an important antecedent to this research, with technologies demonstrated in 1968 subsequently developed further in the interactive computing systems of Xerox PARC. In the early 1980s, Apple will incorporate design ideas associated with Xerox PARC's work on graphical user interfaces and interactive personal computing into its own computers, notably the Lisa and Macintosh.

Historical note: On the legacy of SIMULA

SIMULA emerged from work on simulation and was initially developed in 1962 by Ole-Johan Dahl and Kristen Nygaard as an extension of ALGOL 60. The subsequent development of SIMULA 67 introduced the class concept, which provided a way to organise data and the operations associated with it into reusable structures. These ideas became important in the development of object-oriented programming and influenced later languages, including SMALLTALK.

1970s

1970 MIT's A.I. Group becomes institutionally independent of Project MAC and forms the Artificial Intelligence Laboratory (AI Lab). The Logo Group remains within AI Lab as the Logo Lab, continuing its research and development under Papert's leadership.

1971 Mike Paterson, a visiting computer scientist at MIT, contributes to the development of the LOGO-controlled floor turtle, a physical robot connected to the computer by a cable. Its designation draws inspiration from the cybernetic tortoises developed by Grey Walter. The robot's movements are controlled through Logo instructions, extending programming into physical space.

Papert regards the turtle as a particularly promising educational object. Children can identify with it, imagine themselves in its place, and reproduce its movements bodily, thereby exploring spatial relationships and developing problem-solving strategies (Papert 1980, p. 58). This identification can also help children distance themselves from failure: rather than interpreting an unsuccessful result as a personal failure, they can treat it as a mistake in the turtle's program (Solomon et al. 2020, p. 41).

Papert outlines his vision of A Computer Laboratory for Elementary Schools in Logo Memo No. 1. In Logo Memo No. 2, Teaching Children Thinking, he makes explicit his educational position, drawing on Dewey, Montessori, and Piaget. Together with Solomon, he also publishes Logo Memo No. 3, Twenty Things to Do with a Computer, which proposes a wide range of applications, including activities involving the turtle.

1972 BBN engineer Paul Wexelblat develops Irving, a wireless floor turtle controlled by Logo instructions transmitted by radio. The robot can move, draw, sound its horn, and detect obstacles through contact sensors. Its mobility extends the possibilities for exploring problem-solving strategies, including those involving algebraic relationships (Feurzeig 2011, p. 293).

At MIT, the Logo environment develops in parallel. The computer facilities provide interactive terminals with television-like display screens, alongside printers for hard-copy output (Goldstein et al. 1975, p. 12). A screen turtle, also called a light turtle, is introduced, enabling the movements of an idealised turtle to be represented on a display.

The relationship between physical and screen turtles offers an opportunity to explore isomorphism: the correspondence between the movements of a physical object and their representation through a mathematical model. The physical turtle can enact movements specified through Logo, while the screen turtle provides a visible representation of the corresponding geometric relationships (Papert 1980, p. 56). This relationship between bodily experience, programming, and visual representation becomes an important feature of the Logo educational environment.

1973-1979 Different forms of interaction with the floor turtle are explored. A button-based controller (button box) allows children to direct the turtle without entering commands through the conventional computer interface. The programming of movement can thus be approached independently of the usual Logo interaction.

The physical turtle is also developed with additional sensory capabilities and sound output.

At MIT, the Logo project continues to produce documentation and implementations. Logo Memo No. 7, the Logo Manual, is published in 1974, co-authored by Harold Abelson and collaborators. Abelson, a mathematician and computer scientist, contributes to the development and documentation of the MIT Logo system.

The MIT implementation is written in MACLisp, the Lisp dialect widely used in the AI Lab. Its implementation strategy is documented in Logo Memo No. 11, LLOGO: An Implementation of LOGO in LISP (Goldstein et al. 1974). The report discusses design decisions in the Lisp implementation and compares it with alternative implementations in assembly language. It also documents facilities for character-oriented terminals, graphical turtles, and music generation.

Further applications of LOGO are explored across school subjects, including mathematics—particularly geometry—language education, music, and science. Turtle-based activities, increasingly involving the screen turtle, become a prominent part of this work.

Subsystems of the digital Logo environment designed to support introductory programming, mathematical learning, and musical exploration are investigated in learning laboratories with primary- and secondary-school-aged children.

1976-1977 The Apple Computer 1 (Apple I) is introduced in 1976 as an early microcomputers aimed primarily at computer hobbyists. Its design anticipates the growing availability of relatively affordable computers for individual use. BASIC provides a means of interacting with the machine through a high-level programming language, although the Apple I does not have BASIC preinstalled in ROM.

Apple introduces the Apple II in 1977, opening up further possibilities for personal and educational computing.

In the same year, Terrapin Software is founded to distribute floor turtles. The company subsequently licenses and develops the MIT Logo implementation for the Apple II, helping bring Logo to a wider market.

Historical note: The microcomputer and home-computer era

During the late 1970s and especially the 1980s, increasingly affordable microcomputers bring computing within reach of a much broader public. This development is sometimes described as the home-computer revolution.

The growing availability of personal computers also facilitates the distribution of Logo beyond specialised research environments. Apple becomes an important platform for Logo, and Atari subsequently markets its own Logo implementation. Other systems and software environments follow different paths, with BASIC remaining prominent in many home-computer settings.

The BBN researchers John Seely Brown and Richard R. Burton publish their work on computer-aided diagnosis of errors in arithmetic problem-solving. Their research investigates how a computer can identify patterns in learners' mistakes and use them to support learning (Brown/Burton 1978).

1980s

1980 With Mindstorms: Children, Computers, and Powerful Ideas, Papert presents the work of BBN's Educational Technology Department and MIT's Logo Group to a wider audience. The book develops an educational argument around Logo and the computer as a medium for learning, advancing an emancipatory conception of computer literacy that resonates with Weizenbaum's concerns, and articulating a conception of computational thinking grounded in the expansion of human ways of knowing (Papert 1980, pp. 155, 182; for earlier uses of the term “computational thinking," see Childs 2015; for later interpretations, see Lodi/Martini 2021). Among Mindstorm's central themes are the screen turtle, turtle geometry, and the idea that children can learn by constructing and exploring computational models.

Abelson and Andrea diSessa, a physicist and education researcher, publish Turtle Geometry: The Computer as a Medium for Exploring Mathematics, which focuses on the mathematical ideas that can be explored through turtle geometry. They describe TURTLE TALK as a set of commands that can be implemented as an extension of Logo or adapted to other programming languages.

Smalltalk-80 becomes an influential expression of object-oriented programming and an integrated programming environment. Its approach organises computation around objects that communicate by sending messages. The development of Smalltalk draws on several influences, including Simula, whose class and object concepts also inform other approaches to object-oriented programming. Kay develops the term object-oriented in the course of this work.

Historical note: Different interpretations of object-orientation

Simula's influence also continues along another path. In 1979, the Danish computer scientist Bjarne Stroustrup, working at AT&T Bell Laboratories, begins developing C with Classes, the language that will become C++ in 1984. His aim is to combine facilities for organising programs inspired by Simula with the efficiency and flexibility of C. The resulting language develops along a different path from Smalltalk, reflecting different priorities in programming-language design.

1980 Seymour Papert and several colleagues establish Logo Computer Systems, Inc. (LCSI) to develop and commercialise Logo. Papert becomes chairman, and his former student Brian Silverman becomes director of research. LCSI subsequently develops Apple Logo and versions for other computers, helping Logo reach a wider audience.

Historical note: Another trajectory in the commercialisation of computer technology

In 1976, Bill Gates publishes An Open Letter to Hobbyists, arguing that unauthorised copying of commercial software undermines the economic basis for its development. Microsoft, founded by Gates and Allen in 1975, had begun by selling BASIC for the Altair 8800.

In 1980, Microsoft enters into an agreement with IBM concerning software for its forthcoming personal computer. Microsoft acquires rights to 86-DOS from Seattle Computer Products in 1981, adapts the system for the IBM PC, and licenses it to IBM as PC DOS. Microsoft also distributes its own version as MS-DOS.

86-DOS was strongly influenced by Digital Research's CP/M, an established operating system for earlier microcomputers, but adapted to the Intel 8086 processor. DOS subsequently develops through successive versions, while compatibility with existing software becomes an important concern as IBM-compatible PCs spread.

During the 1980s, IBM-compatible PCs running DOS become increasingly common in German workplaces. Microsoft's software ecosystem, including its later Windows operating systems and Office applications, becomes an important part of this development. This is one strand of a broader transition in which personal computing moves beyond hobbyist and specialist settings into offices and homes.

The development differed across national markets. Apple II had a substantial presence in the United States and an established market in France, whereas Commodore became particularly prominent in West Germany's home-computer market. France also developed a distinctive market shaped by domestic manufacturers, including Thomson, and by the success of Amstrad computers distributed by Schneider. Thus, the spread of personal computing did not follow a uniform international trajectory.

Microsoft's rise was not simply a story of successful products in a diverse and competitive market. As its position strengthened, the company also used aggressive tactics to undermine competing products and reinforce dependence on its own software. A notorious example was the concealed AARD code in beta versions of Windows 3.1, which detected DR-DOS and generated misleading error messages that could lead users to believe that their system was incompatible. DR-DOS was Digital Research's competing operating system, designed to be compatible with MS-DOS and its applications. Internal correspondence revealed that creating uncertainty about DR-DOS was an explicit concern. The episode became part of the legal controversy surrounding Microsoft's competitive practices (Caldera, Inc. v. Microsoft Corp., 72 F. Supp. 2d 1295 (D. Utah 1999)).

1982–1984 Cynthia Solomon leads the newly established Atari Cambridge Research Lab. This group brings together many former members of MIT's Logo Lab and developers from other groups concerned with educational and gaming programming. The lab explores new forms of computer-supported play and learning, combining Logo-based research with ideas drawn from other programming languages, particularly Smalltalk. Its projects include experiments with gesture, music, puppetry, movement, and object-oriented programming.

The St. Paul Logo Project, begun in 1980 in the public schools of St. Paul, Minnesota, continues to investigate the use of Logo in education. Its observations and intermediate findings are published in subsequent reports, including a study published in 1985.

Brian Harvey, a computer science educator at the University of California, Berkeley, and a former member of MIT's Logo Group, publishes the first volume of his series Computer Science Logo Style. Addressed to secondary-school pupils and beginning computer science students, the book introduces programming through symbolic computation, placing the emphasis on programming principles rather than turtle geometry.

1985–1989 Following the closure of the Atari Cambridge Research Lab in 1984, Cynthia Solomon, Margaret Minsky, and Brian Harvey publish LogoWorks: Challenging Programs in Logo in 1985. The collection brings together projects developed by Logo enthusiasts associated with the lab. Its activities range from wordplay, stories, and games to turtle geometry, music, and more advanced programming ideas.

Abelson and diSessa develop Boxer, an experimental computational environment described in their 1986 paper, Boxer: A Reconstructible Computational Medium. In Boxer, semantically meaningful boxes can contain program text, data, and graphical output. Their spatial and hierarchical relationships are part of the computational structure, allowing users to work with programs and their results within an integrated environment rather than treating programming as an activity separate from the interface in which it is carried out.

Logo implementations also explore new ways of representing and controlling multiple turtles. Sprite Logo, developed for the Texas Instruments computers in the late 1970s, had already introduced this possibility. Other implementations extend Logo in different directions. ObjectLOGO, for example, introduces object-oriented programming facilities and lexical scoping, changing how variables and their visibility are handled.

In 1985, the MIT Media Lab is established, incorporating Papert's research into the Epistemology and Learning group. The group develops the educational and theoretical programme of constructionism, investigating how children learn by designing, building, and reflecting on computational objects and environments (Papert, 1986).

In 1986, LogoWriter is released and demonstrated by Papert in a video series. This Logo version extends the dynamic capabilities of the Logo environment text processing, programming and drawing through turtle control.

The LEGO Group's collaboration with the Media Lab grows out of an encounter in 1984 between LEGO owner Kjeld Kirk Kristiansen and Papert, whose ideas about learning through experimentation and play resonate with LEGO's own approach to construction. Beginning in 1985, the collaboration explores how children can program physical constructions using Logo.

Among the participants are Stephen (Steve) Ocko, an educator, filmmaker, and toy developer, and Mitchel (Mitch) Resnick, a computer scientist and doctoral student supervised by Papert, together with other Media Lab colleagues. The resulting LEGO/Logo environment connects Logo programs to constructions made from LEGO elements, including motors and sensors. Children can thus design machines and program their behaviour, extending computational exploration beyond the screen.

Resnick also develops MultiLogo, an agent-based extension of Logo in which multiple turtles can operate independently. The implementation is tested with pupils who already have experience with Logo, providing a setting for exploring programming in terms of interacting agents rather than a single sequence of instructions.

At BBN, Wally Feurzeig and colleagues develop Function Machines, a program that enables the graphical and interactive composition of mathematical functions (Feurzeig, 2011, pp. 298–303). Feurzeig and BBN programmer Karl Troeller also develop the Algebra Workbench, designed to support problem-solving strategies and their symbolic representation. By separating the choice of mathematical operations from the execution of symbolic manipulations, the system allows learners to concentrate on deciding how to proceed while receiving computational assistance with the associated algebraic transformations (Feurzeig, 2011, p. 294).

At Xerox PARC, computer scientists David Ungar and Randall B. Smith develop Self, a programming language that explores another interpretation of object-orientation. Drawing on Smalltalk but adopting a prototype-based object model, Self dispenses with the conventional distinction between classes and instances. Smith's earlier Alternate Reality Kit (ARK) explores related possibilities for constructing and interacting with graphical, programmable worlds.

Work on the Programmable Brick begins at the Media Lab in 1986. The initial design team includes Brian Silverman, Fred G. Martin, Allan Toft, who joins the project from LEGO, Ocko, Resnick, and Papert. A first breadboard prototype is built in 1987. Unlike earlier LEGO/Logo systems, which connect physical constructions to a separate computer, the Programmable Brick aims to incorporate the computer itself into the construction.

The St. Paul Logo Project continues to document and publish its observations of Logo-based learning in schools.

In Costa Rica, the Programa Nacional de Informática Educativa develops an extensive programme of educational computing in public schools. Beginning in the late 1980s, Logo becomes an important tool within the initiative. Teacher education, the preparation of local tutors, and continuing support for participating schools form essential parts of the programme.

In 1988, LEGO introduces LEGO TC Logo, an educational product developed through the collaboration with the Media Lab.

In 1989, LEGO endows a chair at the Media Lab, and Papert becomes the first LEGO Professor of Learning Research.

1990s

1990 The Self programming environment, developed at Xerox PARC, explores prototype-based object-oriented programming and direct manipulation. Its Morphic construction environment, developed by Randall B. Smith, John Maloney, and colleagues, allows graphical objects and programming elements to be arranged and manipulated flexibly. Source text and interface components can be treated as objects within the environment (Smith/Maloney/Ungar 1995a; Maloney/Smith 1995b). Self's prototype-based approach would later be among the influences on JavaScript. Morphic also develops ideas of directness, liveness, and live editing that are relevant to the history of graphical programming environments.

1992 Harvey releases Berkeley Logo, an open-source implementation developed in conjunction with his textbook series, Computer Science Logo Style. In its reference manual, Harvey describes its design as combining “Terrapin-style tokenization” with “LCSI-style syntax,” calling this “the best of both worlds” (Harvey 2020, p. 2). Berkeley Logo thus draws selectively on two established Logo traditions while retaining its own design. It provides a freely available implementation for study, teaching, and further development, but deliberately omits features such as robotics, music, animation, parallelism, and multimedia, which Harvey directs users to commercial implementations for. Berkeley Logo therefore occupies an intermediate position between Terrapin and LCSI in some respects, while all three share a substantial common foundation as Logo implementations.

1993 Logo Computer Systems, Inc. (LCSI) releases MicroWorlds Project Builder, the first product in its MicroWorlds family. It extends the Logo environment with drawing tools, graphics, sound, and animation, supporting the creation of multimedia projects, games, and simulations. The environment builds on the educational ambitions of LogoWriter while taking advantage of more capable personal computers. (LCSI's history of Logo.)

Based on Berkeley Logo, George Mills develops MSWLogo for Microsoft Windows. It makes a textual Logo environment with turtle graphics available on a widely used desktop operating system.

Harvey and Matthew Wright, a researcher in computer music, publish Simply Scheme. The book introduces the Lisp dialect Scheme using a structure and vocabulary adapted from the first two volumes of Harvey's Computer Science Logo Style. It prepares readers for the more advanced treatment of programming in Structure and Interpretation of Computer Programs (SICP), by Abelson, Gerald Jay Sussman, and Julie Sussman (second edition, 1996; online edition). Sussman had been associated with Project MAC and developed Scheme with Guy L. Steele Jr. at MIT. Steele later played a leading role in the development of Common Lisp (Steele et al. 1990).

1994 Resnick advances his work on MultiLogo into StarLogo, a massively parallel version of Logo designed to explore decentralised systems. It enables large numbers of turtles to act independently and interact with one another and with the environment, making it possible to investigate collective behaviour and emergent phenomena (MIT Logo Foundation history).

In collaboration with mathematics education researchers Douglas H. Clements and Julie Sarama, LCSI develops Turtle Math, a version of Logo whose interface and activities are designed to support elementary mathematics learning. Their account of the redesign is published in 1996 (Clements/Sarama 1996).

1995 The project LogoBlocks is started at the MIT Media Lab in connection with the programmable-brick project. LogoBlocks exemplifies a different way of writing programs: users assemble graphical blocks into executable sequences instead of typing lines of textual code. Work associated with Bakhtiar Mikhak, Robby Berg, Martin, and Resnick contributes to this development. The distinction between textual and graphical programming becomes a central question for subsequent educational environments.

1996 Squeak, a new Smalltalk implementation, is released by a team including Ingalls, Kay, Ted Kaehler, Scott Wallace, and Maloney. Its development begins at Apple and continues at Walt Disney Imagineering. Squeak incorporates the Morphic environment, supporting direct manipulation and the creation of interactive objects.

Within this developing ecosystem, Etoys provides a media-rich environment in which learners can draw and animate objects, assign them behaviour, and construct programs through graphical interactions. Its development draws on ideas from Logo and other educational environments. The connections between Squeak, Etoys, and the MIT Media Lab form an important part of the network from which later visual programming environments, including Scratch, develop.

1997 The second edition of Computer Science Logo Style by Brian Harvey appears (contents of Volume 1). The series continues to treat Logo as a language for exploring computer science concepts, including symbolic computation, recursion, higher-order procedures, and programming techniques.

Jim Muller, a Logo enthusiast since 1979, publishes The Great Logo Adventure. The textbook presents graphically oriented teaching activities for primary-school children using MSWLogo and its graphical interface. In its emphasis on visual activities and classroom use, it complements Harvey's more computer-science-oriented treatment of Logo.

1998 Seymour Papert becomes emeritus at MIT. Mitchel Resnick succeeds him as the LEGO Papert Professor of Learning Research at the Media Lab. Drawing inspiration from Friedrich Fröbel and the idea of kindergarten as a model for lifelong learning, Resnick will develop the Lifelong Kindergarten approach to creative learning.

1999 LEGO launches the LEGO Mindstorms Robotics Invention System, featuring the programmable RCX (Robotic Command eXplorer). The product brings programmable robotics, developed through the long-running collaboration between LEGO and the MIT Media Lab, to a much wider audience. The name Mindstorms recalls Papert's influential 1980 book.

Uri Wilensky releases NetLogo, an agent-based modelling environment drawing on StarLogo. It enables users to simulate complex systems through the interactions of many individual agents. Its applications extend beyond elementary programming education to the modelling of phenomena in the natural and social sciences.

2000s

2001 A further version of StarLogo brings the environment to personal computers, continuing the effort to make agent-based modelling accessible beyond specialised research hardware. Subsequent development includes StarLogo TNG, which introduces three-dimensional graphics and a block-based programming interface for creating games and simulations. The last approach will be StarLogo Nova.

2003 Building on earlier work in graphical programming, including LogoBlocks and Etoys, the MIT Media Lab's Lifelong Kindergarten group develops the first prototype of Scratch, implemented in Squeak under Resnick supervision. The development involves Maloney, Jens Mönig, and others. Scratch combines a graphical programming interface with the creation and sharing of interactive stories, games, animations, and other media. Its development brings together ideas and practices from several strands of Logo's history and from related educational computing environments.

2004 David Costanzo continues MSWLogo as the open-source project FMSLogo, extending the life of a textual Logo implementation for Windows.

2006 Jeannette M. Wing, a computer scientist at Carnegie Mellon University in Pittsburgh, publishes “Computational Thinking” in Communications of the ACM. Her article gives influential expression to the idea that concepts and approaches from computer science can inform problem-solving and system design across disciplines. Although Papert had already used the term computational thinking in Mindstorms (1980), Wing's formulation gives it a renewed and distinctive prominence in computer science education and public debate.

Wing completed her bachelor's and master's degrees in electrical engineering and computer science at MIT in 1979 and earned her PhD there in 1983.

2010s

2011 Mönig and Harvey develop BYOB (Build Your Own Blocks), an extension of Scratch that brings features associated with more advanced programming languages into a graphical environment. These include user-defined blocks, recursion, and first-class procedures and data structures. BYOB's development eventually leads to Snap!, which retains Scratch's accessible visual interface while supporting more advanced computer science concepts. Snap! is subsequently reimplemented in JavaScript and develops into an independent project.

2015 The MIT Scratch Team starts Project Microworlds to explore Papert's microworld concept in Scratch.

2020s

2021 Maloney and Mönig are associated with the development of MicroBlocks, a Scratch-like visual programming environment for physical computing. It extends the tradition of accessible graphical programming to microcontrollers and tangible devices, allowing programs to interact directly with sensors, motors, and other hardware.