Friday, August 31, 2018

Turing's journey

A recent post comments on the value of biography as a source of insight into history and thought. Currently I am reading Andrew Hodges' Alan Turing: The Enigma (1983), which I am finding fascinating both for its portrayal of the evolution of a brilliant and unconventional mathematician as well as the honest efforts Hodges makes to describe Turing's sexual evolution and the tragedy in which it eventuated. Hodges makes a serious effort to give the reader some understanding of Turing's important contributions, including his enormously important "computable numbers" paper. (Here is a nice discussion of computability in the Stanford Encyclopedia of Philosophylink.) The book also offers a reasonably technical account of the Enigma code-breaking process.

Hilbert's mathematical imagination plays an important role in Turing's development. Hilbert's speculation that all mathematical statements would turn out to be derivable or disprovable turned out to be wrong, and Turing's computable numbers paper (along with Godel and Church) demonstrated the incompleteness of mathematics. But it was Hilbert's formulation of the idea that permitted the precise and conclusive refutations that came later. (Here is Richard Zack's account in the Stanford Encyclopedia of Philosophy of Hilbert's program; link.)

And then there were the machines. I had always thought of the Turing machine as a pure thought experiment designed to give specific meaning to the idea of computability. It has been eye-opening to learn of the innovative and path-breaking work that Turing did at Bletchley Park, Bell Labs, and other places in developing real computational machines. Turing's development of real computing machines and his invention of the activity of "programming" ("construction of tables") make his contributions to the development of digital computing machines much more advanced and technical than I had previously understood. His work late in the war on the difficult problem of encrypting speech for secure telephone conversation was also very interesting and innovative. Further, his understanding of the priority of creating a technology that would support "random access memory" was especially prescient. Here is Hodges' summary of Turing's view in 1947:
Considering the storage problem, he listed every form of discrete store that he and Don Bayley had thought of, including film, plugboards, wheels, relays, paper tape, punched cards, magnetic tape, and ‘cerebral cortex’, each with an estimate, in some cases obviously fanciful, of access time, and of the number of digits that could be stored per pound sterling. At one extreme, the storage could all be on electronic valves, giving access within a microsecond, but this would be prohibitively expensive. As he put it in his 1947 elaboration, ‘To store the content of an ordinary novel by such means would cost many millions of pounds.’ It was necessary to make a trade-off between cost and speed of access. He agreed with von Neumann, who in the EDVAC report had referred to the future possibility of developing a special ‘Iconoscope’ or television screen, for storing digits in the form of a pattern of spots. This he described as ‘much the most hopeful scheme, for economy combined with speed.’ (403)
These contributions are no doubt well known by experts on the history of computing. But for me it was eye-opening to learn how directly Turing was involved in the design and implementation of various automatic computing engines, including the British ACE machine itself at the National Physical Laboratory (link). Here is Turing's description of the evolution of his thinking on this topic, extracted from a lecture in 1947:
Some years ago I was researching on what might now be described as an investigation of the theoretical possibilities and limitations of digital computing machines. I considered a type of machine which had a central mechanism and an infinite memory which was contained on an infinite tape. This type of machine appeared to be sufficiently general. One of my conclusions was that the idea of a ‘rule of thumb’ process and a ‘machine process’ were synonymous. The expression ‘machine process’ of course means one which could be carried out by the type of machine I was considering…. Machines such as the ACE may be regarded as practical versions of this same type of machine. There is at least a very close analogy. (399)
At the same time his clear logical understanding of the implications of a universal computing machine was genuinely visionary. He was evangelical in his advocacy of the goal of creating a machine with a minimalist and simple architecture where all the complexity and specificity of the use of the machine derives from its instructions (programming), not its specialized hardware.

Also interesting is the fact that Turing had a literary impulse (not often exercised), and wrote at least one semi-autobiographical short story about a sexual encounter. Only a few pages survive. Here is a paragraph quoted by Hodges:
Alec had been working rather hard until two or three weeks before. It was about interplanetary travel. Alec had always been rather keen on such crackpot problems, but although he rather liked to let himself go rather wildly to newspapermen or on the Third Programme when he got the chance, when he wrote for technically trained readers, his work was quite sound, or had been when he was younger. This last paper was real good stuff, better than he'd done since his mid twenties when he had introduced the idea which is now becoming known as 'Pryce's buoy'. Alec always felt a glow of pride when this phrase was used. The rather obvious double-entendre rather pleased him too. He always liked to parade his homosexuality, and in suitable company Alec could pretend that the word was spelt without the 'u'. It was quite some time now since he had 'had' anyone, in fact not since he had met that soldier in Paris last summer. Now that his paper was finished he might justifiably consider that he had earned another gay man, and he knew where he might find one who might be suitable. (564)
The passage is striking for several reasons; but most obviously, it brings together the two leading themes of his life, his scientific imagination and his sexuality.

This biography of Turing reinforces for me the value of the genre more generally. The reader gets a better understanding of the important developments in mathematics and computing that Turing achieved, it presents a vivid view of the high stakes in the secret conflict that Turing was a crucial part of in the use of cryptographic advances to defeat the Nazi submarine threat, and it gives personal insights into the very unique individual who developed into such a world-changing logician, engineer, and scientist.

Wednesday, August 29, 2018

The insights of biography

I have always found biographies a particularly interesting source of learning and stimulation. A recent example is a biography and celebration of Muthuvel Kalaignar Karunanidhi published in a recent issue of the Indian semi-weekly Frontline. Karunanidhi was an enormously important social and political leader in India for over sixty years in the Dravidian movement in southern India and Tamil Nadu, and his passing earlier this month was the occasion for a special issue of Frontline. Karunanidhi was president of the Dravidian political party Dravida Munnetra Kazhagam (DMK) for more than fifty years. And he is an individual I had never heard of before opening up Frontline. In his early life he was a script writer and film maker who was able to use his artistic gifts to create characters who inspired political activism among young Tamil men and women. And in the bulk of his career he was an activist, orator, and official who had great influence on politics and social movements in southern India. The recollection and biography by A.S. Panneerselvan is excellent. (This article derives from Panneerselvan's forthcoming biography of Karunanidhi.) Here is how Panneerselvan frames his narrative:
In a State where language, empowerment, self-respect, art, literary forms and films coalesce to lend political vibrancy, Karunanidhi's life becomes a sort of natural metaphor of modern Tamil Nadu. His multifaceted personality helps to understand the organic evolution of the Dravidian Movement. To understand how he came to the position to wield the pen and his tongue for his politics, rather than bombs and rifles for revolution, one has to look at his early life. (7)
I assume that Karunanidhi and the Dravidian political movement would be common currency for Indian intellectuals and political activists. For an American with only a superficial understanding of Indian politics and history, his life story opens a whole new aspect of India's post-independence experience. I think of the primary dynamic of Indian politics since Independence as being a struggle between the non-sectarian political ideas of Congress, the Hindu nationalism of BJP, and the secular and leftist position of India's Communist movement. But the Dravidian movement diverges in specific ways from each of these currents. In brief, the central thread of the Dravidian is the rejection of the cultural hegemony of Hindi language, status, and culture, and an expression of pride and affirmation in the cultures and traditions of Tamil India. Panneerselvan describes an internal difference of emphasis on the topic of language and culture within the early stage of the Dravidian movement:
The duality of the Self-Respect Movement emerged very clearly during this phase. While Periyar and Annadurai were in total agreement in the diagnosis of the social milieu, their prognoses were quite opposite: For Periyar, language was an instrument for communication; for Annadurai, language was an organic socio-cultural oeuvre that lends a distinct identity and a sense of pride and belonging to the people. (13).
The Dravidian Movement was broadly speaking a movement for social justice, and it was fundamentally supportive of the rights and status of dalits. The tribute by K. Veeramani expresses the social justice commitments of DMK and Karunanidhi very well:
The goal of dispensation of social justice is possible only through reservation in education and public employment, giving adequate representation to the Scheduled Castes, the Scheduled Tribes and Other Backward Classes. Dispensation of social justice continues to be the core principle of the Dravidian movement, founded by South Indian Liberal Federation (SILF), popularly known as the Justice Party. (36) ... The core of Periyar's philosophy is to bring about equality through equal opportunities in a society rife with birth-based discrimination. Periyar strengthened the reservation mode as a compensation for birth-based inequalities. In that way, reservation has to be implemented as a mode of compensatory discrimination. (38)
Also important in the political agenda of the Dravidian Movement was a sustained effort to improve the conditions of tenants and agricultural workers through narrowing of the rights of landlords. J. Jeyaranjan observes:
The power relation between the landlord and the tenant is completely reversed, with the tenant enjoying certain powers to negotiate compensation for giving up the right to cultivate. Mobilisations by the undivided Communist Party of India (CPI) and the Dravidian movement, the Dravidar Kazhagam in particular, have been critical to the creation of a culture of collective action and resistance to landlord power. Further, the coming to power of the Dravida Munnetra Kazhagam (DMK) in 1967 created conditions for consolidating the power of lower-caste tenants who benefited both from a set of State initiatives launched by the DMK and the culture of collective action against Brahmin landlords. (52)
What can be learned from a detailed biography of a figure like Karunanidhi? For myself the opportunity such a piece of scholarship permits is to significantly broaden my own understanding of the nuances of philosophy, policy, values, and institutions through which the political developments of a relatively unfamiliar region of the world have developed. Such a biography allows the reader to gain a vivid experience of the issues and passions that motivated people, both intellectuals and laborers, in the 1920s, the 1960s, and the 1990s. And it gives a bit of insight into the complicated question of how talented individuals develop into impactful, committed, and dedicated leaders and thinkers.

(Here is a collection of snippets from Karunandhi's films; link.)

Sunday, August 19, 2018

Safety culture or safety behavior?

Andrew Hopkins is a much-published expert on industrial safety who has an important set of insights into the causes of industrial accidents. Much of his career has focused on the oil and gas industry, but he has written on other sectors as well. Particularly interesting are several books: Failure to Learn: The BP Texas City Refinery Disaster; Disastrous Decisions: The Human and Organisational Causes of the Gulf of Mexico Blowout; and Lessons from Longford: The ESSO Gas Plant Explosion. He also provides a number of interesting working papers here.

One of his interesting working papers is on the topic of safety culture in the drilling industry, "Why safety cultures don't work" (link).
Companies that set out to create a “safety culture” often expend huge amounts of resource trying to change the way operatives, foremen and supervisory staff think and feel about safety. The results are often disappointing. (1)
Changing the way people think is nigh impossible, but setting up organizational structures that monitor compliance with procedure, even if that procedure is seen as redundant or unnecessary, is doable. (3)
Hopkins' central point is that safety requires change of routine behavior, not in the first instance change of culture or thought. This means that management and regulatory agencies need to establish safe practices and then enforce compliance through internal and external measures. He uses the example of seat belt usage: campaigns to encourage the use of seat belts had little effect, but behavior changed when fines were imposed on drivers who continued to refrain from seat belt usage.

His central focus here, as in most of his books, is on the processes involved in the drilling industry. He makes the point that the incentives that are established in oil and gas drilling are almost entirely oriented towards maximizing speed and production. Exhortations towards "safe practices" are ineffectual in this context.

Much of his argument here comes down to the contrast between high-likelihood, low-harm accidents and low-likelihood, high-harm accidents. The steps required to prevent low-likelihood, high-harm accidents are generally not visible in the workplace, precisely because the sequences that lead to them are highly uncommon. Routine safety procedures will not reduce the likelihood of occurrence of the high-harm accident.

Hopkins offers the example of the air traffic control industry. The ultimate disaster in air traffic control is a mid-air collision. Very few such incidents have occurred. The incident Hopkins refers to was a mid-air collision over Uberlinger, Germany in 2002. But procedures in air traffic control give absolute priority to preventing such disasters, and the solution is to identify a key precursor event to a mid-air collision and ensure that these precursor events are recorded, investigated, and reacted to when they occur. The relevant precursor event in air traffic control is a proximity of two aircraft at a distance of 1.5 miles or less. The required separation is 2 miles. Air traffic control regulations and processes require a full investigation and reaction for all incidents of separation that occur with 1.5 miles of separation or less. Air traffic control is a high-reliability industry precisely because it gives priority and resources to the prevention, not only of the disastrous incidents themselves, but the the precursors that may lead to them. "This is a clear example of the way a high-reliability organization operates. It works out what the most catastrophic event is likely to be, regardless of how rare such events are in recent experience, and devises good indicators of how well the prevention of that catastrophe is being managed. It is a way of thinking that is highly unusual in the oil and gas industry" (2).

The drilling industry does not commonly follow similar high-level safety management. A drilling blowout is the incident of greatest concern in the drilling industry. There are, according to Hopkins, several obvious precursor events to a well blowout: well kicks and cementing failures. It is Hopkins' contention that safety in the drilling industry would be greatly enhanced (with respect to the catastrophic events that are both low-probability and high-harm) if procedures were reoriented so that priority attention and tracking were given to these kinds of precursor events. By reducing or eliminating the occurrence of the precursor events, major accidents would be prevented.

Another organizational factor that Hopkins highlights is the role that safety officers play within the organization. In high-reliability organizations, safety officers have an organizationally privileged role; in low-reliability organizations their voices seem to disappear in the competition among many managerial voices with other interests (speed, production, public relations). (This point is explored in an earlier post; link.)
Prior to Macondo [the Deepwater Horizon oil spill], BP’s process safety structure was decentralized. The safety experts had very little power. They lacked strong reporting lines to the centre and answered to commercial managers who tended to put production ahead of engineering excellence. After Macondo, BP reversed this. Now, what I call the “voices of safety” are powerful and heard loud and clear in the boardroom. (3)
Ominously, Hopkins makes a prescient point about the crucial role played by regulatory agencies in enhancing safety in high-risk industries.
Many regulatory regimes, however, particularly that of the US, are not functioning as they ought to. Regulators need to be highly skilled and resourced and must be able to match the best minds in industry in order to have competent discussions about the risk-management strategies of the corporations. In the US they're not doing that yet. The best practice recognized worldwide is the safety case regime, in use in UK and Norway. (4)
Given the militantly anti-regulatory stance of the current US federal administration and the aggressive lack of attention its administrators pay to scientific and technical expertise, this is a very sobering source of worry about the future of industrial, chemical, and nuclear safety in the US.