Algorithmic Language - Cuba
"Both the revolutionaries and their enemies recognized that information technology would be a strategic priority for the new Cuba. A former IBM executive recalls that “all of the foreign enterprises had been nationalized except for IBM Cuba,” since the “Castro government and most of the nationalized companies were users of IBM equipment and services.” But from 1961–62, IBM closed its Cuban branch, and the US government imposed a trade embargo that prevented Cuba from acquiring computer equipment. This meant that Cuba would be forced to develop its own computing industry, with help from other socialist countries in the Soviet-led Council for Mutual Economic Assistance (Comecon).
Between 1969 and 1970, a team at the University of Havana created a prototype of a digital computer, the CID-201, as well as an assembly language named LEAL, short for “Lenguaje Algorítmico” (Algorithmic Language), an acronym that also means “loyal.” The design of the CID-201 was based on the schematics found in the manual of the PDP-1, a computer manufactured by the US-based Digital Equipment Corporation. Because of the US-imposed trade embargo, the team could not buy the necessary electronic components in Europe, but eventually succeeded—with the help of a Cuban man of Japanese descent who worked as a merchant in Tokyo—in bringing the components from Japan inside more than ten briefcases.
Cuban mathematicians also wrote a computer program in LEAL for playing chess; one of the CID-201’s engineers recounts that the computer even played—and lost—a game against Fidel Castro. Starting in the 1970s, Cuba manufactured thousands of digital computers, and even exported some computer parts to other Comecon countries.
The rise of digital computing transformed Cuban librarianship. Freyre de Andrade welcomed the digital age, paraphrasing Marx and Engels to analogize computing to communism: “a specter is haunting the informational world, the specter of the computer; and let’s be pleased that this circumstance has come to move our field [of librarianship], giving us a challenge that makes [the field] even more interesting than it already was by itself.” Cubans studied the techniques of informatics mostly with Soviet textbooks translated into Spanish. They combined the computational methods they learned from these books with the revolutionary ideals of Cuban librarianship. This synthesis produced distinctive theories and practices that diverged substantially from those of both Western and Soviet informatics." (https://logicmag.io/care/informatics-of-the-oppressed/?)
"Consider the concept of “information laws,” a staple of informatics textbooks. A classic example is “Lotka’s law,” formulated in 1926 by Alfred J. Lotka, a statistician at Metropolitan Life Insurance Company in New York, who sought to compute the “frequency distribution of scientific productivity” by plotting publication counts of authors included in an index of abstracts of chemistry publications. He claimed that the distribution followed an “inverse square law,” i.e., “the number of persons making 2 contributions is about one-fourth of those making one; the number making 3 contributions is about one-ninth, etc.; the number making n contributions is about 1/n² of those making one.”
Like Western textbooks, the Soviet textbooks of informatics adopted in Cuba covered such “information laws” in depth. Their main authors, Russian information scientists and engineers A. I. Mikhailov and R. S. Gilyarevskii, quoted a peculiar passage by US information scientist and historian of science Derek de Solla Price on the distribution of publication counts: “They follow the same type of distribution as that of millionaires and peasants in a highly capitalistic society. A large share of wealth is in the hands of a very small number of extremely wealthy individuals, and a small residual share in the hands of the large number of minimal producers.”
For Cuban information scientists, who had experienced a socialist revolution and an abrupt redistribution of material wealth, this unequal distribution of informational wealth also had to be radically transformed. Among these information scientists was Emilio Setién Quesada, who had studied and worked with Freyre de Andrade since the beginning of the post-revolutionary period. Setién Quesada contested the very idea of an “information law.” In an article co-authored with a Mexican colleague, he objected to the term “law,” which seemed to imply “the identification of a causal, constant, and objective relation in nature, society, or thought.” The mathematical equations represented mere “regularities,” without expressing “the causes of qualitative character of the behaviors they describe.” Those causes were historical, not natural.
Therefore, Setién Quesada and his colleague argued, publication counts did not conclusively determine the “productivity” of authors, any more than declining citation counts indicated the “obsolescence” of publications. Cuban libraries shouldn’t rely on these metrics to make such consequential decisions as choosing which materials to discard. Traditional informatics was incompatible with revolutionary librarianship because, by treating historically contingent regularities as immutable laws, it tended to perpetuate existing social inequalities.
Cuban information scientists didn’t just critique the limitations of traditional informatics, however. They also advanced a more critical approach to mathematical modeling, one that emphasized the social complexity and the historical contingency of informational regularities. In the 1980s, when Cuban libraries were beginning to adopt digital computers, Setién Quesada was tasked with developing a mathematical model of library activity, based on statistical data, for the purpose of economic planning. But he was dissatisfied with existing models of the “intensity” and “effectiveness” of library activity, devised by Soviet and US information scientists. (In the discussion below, I include mathematical explanations inside parentheses for interested readers, following Setién Quesada’s own terminology and notation.)
Soviet information scientists computed the “coefficient of intensity” of library activity by multiplying the “index of circulation” (the number of borrowings m divided by the number of potential readers N) by the “index of rotation” (the number of borrowings m divided by the total volume of holdings f). Meanwhile, US information scientists computed the “measure of effectiveness” of libraries, combining the index of circulation with an “index of capture” (the number of actual library readers n divided by the number of potential readers N). In contrast to these two approaches, Setién Quesada proposed an alternative “Cuban model,” which evaluated what he called the “behavior of Cuban public libraries”.
Setién Quesada argued that “the Cuban model is more complete.” It included many more variables, all of which he considered important. For instance, the Cuban model included an “index of communication” (based on the number l of readers who use the archive), while the Soviet and US models “do not express the precise level of the author-reader social communication that happens in libraries.” Moreover, those other models “do not consider the role of the librarian in the development of the activity.” For Setién Quesada, the librarians, “together with the readers, constitute the main active agents involved in the development of this activity.” Hence in the Cuban model, every variable was adjusted relative to the number of librarians (incorporated into the adjusted variables denoted by a vinculum). Finally, the other models “do not offer an index that synthesizes the comparative behavior of places and periods.” By contrast, the Cuban model sought to facilitate comparisons of different libraries and time periods (each represented by the subscript i).
Whatever the merits and limitations of this particular mathematical model, the broader story of Cuban information science encourages us to be skeptical of the claims attached to models and algorithms of information retrieval in the present. If yesterday’s information scientists claimed that their models ranked authors by “productivity” and libraries by “effectiveness,” today’s “AI experts” claim that their algorithms rank “personalized” search results by “relevance.” These claims are never innocent descriptions of how things simply are. Rather, these are interpretive, normative, politically consequential prescriptions of what information should be considered relevant or irrelevant.
These prescriptions, disguised as descriptions, serve to reproduce an unjust status quo. Just as print publications should not be deemed obsolete and discarded from library collections on the basis of citation counts, online information should not be deemed irrelevant and ranked low in search results on the basis of “click-through rates” and ad revenues. The innovative experiments by Cuban information scientists remind us that we can design alternative models and algorithms in order to disrupt, rather than perpetuate, patterns of inequality and oppression." (https://logicmag.io/care/informatics-of-the-oppressed/?)