by Lev Lafayette
ISBN-10:
ISBN-13: 978-0-9943373-5-1
Teaching Research Computing to Advanced Learners by Lev Lafayette, 2015
Published by the Victorian Partnership for Advanced Computing (trading as V3 Alliance) .
Cover art composed by Michael D'Silva, featuring several clusters operated by the Victorian Partnership for Advanced Computing.
Teaching Research Computing to Advanced Learners, is licensed under a Creative Commons Attribution 4.0 International License.
All trademarks are property of their respective owners.
0.0 Introduction 0.1 Foreword 0.2 Preface
1.0 Context 1.1 The Knowledge Economy 1.2 The Adult Learner 1.3 Lifelong Learning
2.0 Audience 2.1 Advanced Andragogy 2.2 Learner Motivation 2.3 Audience Diversity
3.0 Content 3.1 Learning Objectives 3.2 Structured Content 3.3 Proximal Positioning 3.4 Provisioning
4.0 Delivery 4.1 Planning 4.2 Disciplinary Learning Styles 4.3 Two Types of Constructs 4.4 Collaboration and Connectivism
5.0 Assessment and Revision 5.1 Course Evaluation 5.2 Course Revision 5.3 Learner Assessment
6.0 Appendicies
Without giving too much away - and it should be evident from the title - this book is for the teaching of particular computing skills for researchers that are primarily in non-computer science fields. It is borne from many years of experience in teaching computing to postgraduate and postdoctoral researchers to a number of institutions across Australia and can be accompanied by a number of books also recently published which derive their content from those classes. These publications include: Supercomputing with Linux, Sequential and Parallel Programming, Mathematical Applications and Programming, and Software Tools for eResearchers. A great deal of the motivation for writing this book comes from the need to 'teach the teachers', for as many a computer science student will affirm, those who have excellent technical knowledge are not necessarily those who have excellent educational knowledge. This is by no means a slight on their part, but rather simply a recognition that they are different skillsets, and one does not imply the other.
Another motivation in writing the publication is being witness to the negative effects of the improvements in computer interfaces since the early 1990s. Whilst there can be little doubt in the gains in providing intuitive user interfaces in simplifying a number of tasks and by lowering the knowledge requirements to actually engage in computer operations, there has been a loss in the general level of understanding among computer users. Those who were raised with the command-line as the primary interface have a better understanding of file system hierarchies and often operating system components, than those who have only gone to such levels under duress. The problem becomes evident however where, under circumstances where they have massive data processing needs, that high performance computation is required and the necessary skills.
Many authors are often at pains to avoid expressing moral judgements in their works and seek to remain as objective as possible. Whilst objectivity is very much a laudable goal, taking ethical positions themselves is by no means a matter of conflict. Indeed, in terms of evaluation it is clear that they are incommensurable. Facts may provide the resources for ethical positions, but they cannot determine them. Ethical positions require a grounding in evidence in order to be effective, but cannot determine the evidence. The pragmatic combination of two - what the ancient Hellenes called phronesis - combines both resources in a manner that results in prudent action. Thus, rather than being "just" factual, this publication finds inspiration from educational philophers whose commitment is towards an informed democratic participation and can include figures such as John Dewey, Jürgen Habermas, and the emancipatory orientations of Paulo Freire. In terms of developmental psychology and education, primary influences are Jean Piaget and Lawrence Kohlberg, and Malcolm Knowles. On a related topic one other matter than must be mentioned as part of an democratic educational philosophy is a commitment to open-source knowledge. This commitment is based on the premise that research is best conducted when it is available for peer review.
Knowing how to use a computer is the equivalent of literacy of our age, and knowing how to use it with a higher degree of competency is essential for researchers. Those who do not have these skills are at a distinct disadvantage. They cannot navigate to the extensive sources of knowledge available online, they cannot access the enormous data stores, they cannot process that data to make new discoveries. Certainly such individuals may also be protected from the rapid distribution of hyperbolic false information that appeals to confirmation and selective biases. But it is even more possible that in the connected enviroment that individuals will encounter those who have contrary views and information. Depending on their disposition they may even change their opinions when provided new evidence, which is certainly a prequisite for a genuine learner. Whilst people may learn from confirmation of their existing presumptions, but even more so from the intellectual humility to recognise their own ignorance on a subject. Reading a number of online 'discussions', one does wonder of the relative distribution of the two.
Of course, before this happens the user must have the access to the systems and competency in their use. Whilst the former is beyond the scope of this book, being more of a subject for sociologists and economists, the latter is the central issue here of which there are several components. Firstly, there is the question of what competencies are to be imparted, starting from the high-level discipline and then into the subdisciplines and components. Much of this is outside of the control of the educator themselves and depends more on the social needs of the learner and the determinations of the institution which the teaching is being conducted. The first chapter of this publication will initially discuss this in in some detail, as it also determines the disciplinary competencies and processes that the educator should concentrate on. After all, in order to impart knowledge the educator must have the knowledge themselves, notwithstanding the specific skills for education, or the new knowledge that can be gained in the process of conducting education.
The second chapter discusses the audience that an educator is involved with. The reality is that different learners are have differing levels of experience and different levels of cognitive development. In one sense, nearly all educators realise that it is simply not possible to apply the same content to different levels of experience; one does not use the same content in a programming course to upper primary school students as they would to post-graduate researchers. This much is obvious - what is much more surprising is that some educators seem to think that the same approach can be used in such divergent audiences. In this manner the conceptus of the pedagogy-andragogy continuum is illustrated, and the particular characteristics of adult learners, and especially advanced adult learners. Even if they have the same structures of cognitive development, the post-doctoral researcher is a different learner to the first-year undergratuate. In a different manner, the educator needs to be sensitive to the different experiences and meanings that a learner has based on the intersectional diversities of cultural, religious, gender, and class backgrounds. Understanding how to to harness these experiences in a constructive manner can further aid the learner by appealing to core issues of their social identity.
With the context and audience largely identified, the educator can then develop and emphasise appropriate content. The first step of this process is really establishing the learning objectives of the content, which itself is often part of the jigsaw of a wider programme or degree, which also indicates the prerequisites, if any, of the educational package. Any outcomes can then be used as prerequisites for other training packages. Within each training package however there needs to be structure to the content. This certainly does not mean that there is provisioning of style headings in a document, as one rather unfortunate bureaucrat in search of meaningful employment once tried to suggest (OK, granted that helps). Rather structured content, in an educational sense, is the application of input prequisites and output competencies within the learning module itself. If all of this sounds like procedural programming and the passing of parameters that's certainly because there is a high degree of similarity!
In addition to learning objectives and structure, the content itself also requires to be attentive to learning proximity. What is meant by this is the (originally) pedagogical theory of Lev Vygtotsky, who argued that for a particular compentency between independent competence and incompetence, there is a area ("the zone of proximal development") where a learner can engage with a task where they have an assistant or guide to help them through the process. It is in this zone that content should be primarily developed and where the teacher can provide the most practical assistance. Beyond this, there is the question of provisioning of the content. That is, what content should be provided and where. This may come with institutional restrictions, but in terms of general educational benefits free and open provisioning is explored.
After developing the core content, there is the matter of delivery. In an environment where there is face-to-face interaction with a teacher, this will almost invariably lead to the introduction ad-hoc content, especially in an advanced andragogical audience. Nevertheless this does not diminish the need for detailed lesson plans. With these in place, a particularly disciplinary learning style should be adopted. There has been a lot of justified criticism of the suggestion of individual learning styles, however there is empirical evidence that learning styles do have influence according to the subject matter. In the field of computer education, the principles of "yield to the hands-on imperative" (as the famous 'Hackers: Heroes of the Computer Revolution" put it) and paired-programming, provide a useful foundation. These contribute to constructivism, the educational theory from Jean Piaget that people generate knowledge and meaning from the interaction between their abstract ideas and concrete experiences, and constructionism, Seymour Papert's argument that learners construct intellectual models in the pursuit of understanding their experiences and the ideas provided to them. In this particular discipline, much is to be gained from the new 'connectivist' approaches to content delivery, inspired by contemporary social media technologies.
After delivery of a learning package assessment and revision is absolutely essential. Obviously, there should be some form of assessment of the learner's as well. However in an advanced andragogical environment, traditional forms of assessment are not particularly well suited. Alternatives which are based on actual use of the material are far superior for this particular audience. With regards to revision, nobody's teaching material is perfect, least of all this publication, and nobody's delivery is perfect either. One should always seek to improve and update their material. This includes the actual content itself, the delivery, and the facilities where the teaching is conducted. It is argued that the first body of opinion generation should come from the learners themselves.
The content of this book is largely derived from the experience of training postgraduate researchers through the Victorian Partnership for Advanced Computing. These classes were designed to provide the necessary skills for such learners to utilse the Linux operating system effectvely (navigation, shell scripting, Python scripting), use various batch-system job submissions tools for high performance compute clusters (TORQUE, PBSPro, OpenPBS, write parallel programs (OpenMPI), use mathematical programming languages (R, Octave), and database systems (PostgreSQL), along with training courses for system administrators in managing their job schedulers, resource managers, distributed compute management tools (xCAT), monitoring systems (Ganglia, Nagios), and compiling various scientific software applications.
Such training was carried out over a number of institutions including: the Royal Melbourne Institute of Technology, La Trobe University, the University of Melbourne, Deakin University, Swinburne University, Victoria University of Technology, the University of Sydney, Macquarie University, the University of New South Wales, the University of Western Australia, the Australian Institute of Health Innovation, the Westmead Millennium Institute, and the Australian Institution of Marine Science. Prior I also had experience training staff in general computer systems at the Ministry of Foreign Affairs and Cooperation in Timor-Leste, and several years training the electorate office staff at the Parliament of Victoriatraining electoral roll database systems.
One of the significant challenges for this book was to make it generic enough so it could be applied for a range of computer technologies but also practical enough so that a computer science educator would not be bogged down in just theory, fascinating enough as it is. Hopefully one of the main aspects of advanced education - that is to provide theoretically grounded practical skills - has been delivered, albeit only in this particular format. The publication is partially supplemented from a variety of papers and presentations from the author to a number of eResearch educational conferences and other similar groups over the past several years, but more directly from over a thousand hours of classes to researchers over the past eight years. It was their feedback and enthusiasm that provided the greatest motivation for publication.
Lev Lafayette, Melbourne 2015
It may seem strange to begin a book on adult education and research computing with an excursis into the field of theoretical macroeconomics, but nevertheless it is grounded in practical intent. Before one engages in teaching, the question must be asked 'what is this teaching for?' and whilst the trite answer 'because there is a demand for it', is satisfactory from a producer-consumer relationship, the wider implications ("externalities") must also be addressed. Transactions, including the transfer of knowledge and skills, have additional effects on others who are not part of that transaction.
In the past several decades it has been fashionable to discuss individual and aggregate knowledge and skills in terms of human capital. The OECD has defined human capital as follows:
"The knowledge, skills, competencies and attributes embodied in individuals that facilitate the creation of personal, social and economic well-being" (OCED, 'The Well Being of Nations', 2011).
In this sense, education and training is seen as an 'investment' in the learner, which supplements their innate qualities and other environmental factors, and one which should have the outcomes of higher earnings, productivity gains, individual satisfaction, and social integration. The latter point is an excellent example of positive externlities. Additional participation in education has been shown to correlate with a diverse range of benefits including better health outcomes, reduced risk of smoking, reduced criminal involvement, and higher participation in community groups (OECD, op cit). These correlations are certainly not reducible to cognitive ability alone.
Variations in individual earnings are partially explicible by human capital investments as well. The Interantional Adult Literacy Survey (IALS) accounted for some 40% of variation in individual earnings through measures such as education, literacy and work experience, parental educational attainment, and so forth. Interestingly of course is that some 60% can not be attributed to such factors.
In a model of perfect competition, wages are determined by productivity. However labour markets do not operate in perfect competition; in most cases they are monopsonistic, that is, there are more potential employees than employers, giving market power to the latter group. In addition there is conscious and unconscious discrimination by employers according to social categories; bordering on the bizarre women orchestra players have been traditionally underrepresented with a range of excuses provided concerning competence. When 'blind auditions' where conducted the likelihood of female selection increased by sevenfold (Marilyn Marks, Blind auditions key to hiring musicians, Princeton Weekly Bulletin, Feb 12, 2001 http://www.princeton.edu/pr/pwb/01/0212/7b.shtml). Nevertheless, it is clear that education and training does have positive individual and social outcomes.
There is also a semantic and category debate as well. In political economy, capital is what is owned by capitalists, from which they receive an income return from. It is machinery part of the factors of production, distinct from natural resources and labour. The confusion between material capital and considering human beings as capital by which they can be invested in and owned, can lead to something akin to slavery:
"The counterpart for education would be to 'buy' a share in an individual's earning prospects: to advance him the funds needed to finance his training on condition that he agree to pay the lender a specified fraction of his future earnings. In this way, a lender would get back more than his initial investment from relatively successful individuals, which would compensate for the failure to recoup his original investment from the unsuccessful. There seems no legal obstacle to private contracts of this kind, even though they are economically equivalent to the purchase of a share in an individual's earning capacity and thus to partial slavery." (Milton Friedman, Capitalism and Freedom, 1962)
In classic political economy, the primary factors of production were land (i.e., all natural resources), labour, and capital. Capital represents durable goods that are used in production of goods or services. It contrasts with the accounting definition of "capital" which refers to stocks, shares, currency etc representing the perceived exchange value of such durable goods. There is a belief that political economy did not include "intangibles", such as organisation efficiencies, individual entrepreneurship, consumer goodwill, or skills and knowledge. This isn't quite true, as there was recognition for example, of "labour power", the variable capacity of each additional unit to the productive process, or likewise the productivity of machinery, or the from various grades of land (from Ricardo).
Nevertheless the perception of this alleged lack has led to a proliferation of capital theory. Today there is volumes of material describing "human capital" (i.e., skills and knowledges), "social capital" (networks of trust), "intellectual capital" (intellectual property), "natural capital" (the environment), even "cultural capital" (particular aesthetic tastes) and so forth. How long will it be until someone produces a treatise of "capital capital"?
Whilst one could engage in a metaphor of investing in knowledge and skills (i.e., education and training) to produce utility ("well being") and that these knowledge and skills may depreciatate, and could be subject to "capital flight" as 'owners' of the human capital depart for places that offer better wages. If the metaphor is followed further perhaps "attributes" could be considered something closer to "human land" instead, although it should be pointed out that neoclassical economics tends to, incorrectly in my view, conflate land and capital. Indeed arguably, the discussion of "human capital" does look like an attempt to convert labour, as a factor production, into capital as well. Apparently everything is capital now!
From the outset it seems that the term "human capital" is not really the right one, as it does not share the definitional characteristics of "capital" in an economic sense, and nore does it share the ownership relations. The same can be said for the other It would be better to call investment in labour what it actually is - and remove an potential ambiguity that the owner of skills and knowledges (and attributes and motivations), is the individual where they are actually located. Where such terms are encountered, it is more useful to think of "training and education" for the purpose of improving potential well-being.
The term "potential" is used carfully. Phillip Brown (P. Brown, 'The Opportunity Trap: Education and Employment ina Global Economy', European Education Research Journal, Vol 2:1, 2003, p141-179) explictly takes issues with the claims from the OECD with the claims that investment in education provides "opportunity, prosperity[,] and justice", specifically arguing that "social inequalities in the competition for a livelihood and an intensification of 'positional' conflict". Everyone must engage in this rush for positional advantage in order to maintain their current standard of living. We "have to run faster, for longer, just to stand still". This is "the opportunity trap".
The argument elaborates both prior symptoms of the "diploma disease" and uses the core insights of the sociology of education as elucidated by Durkheim (i.e., the socialisation role, and the selection role), and Weber (rationalisation of society, complexity). There are claims that the new professions generated by technical complexity will "generate new opportunities for jobs that offer excitement, creative fulfillment, and personal development. Work is the new consumption." From a conflict perspective however, "The expansion of higher education is not believed to reflect changes in the demand for high-skilled workers, but credential inflation ... The idea of credential inflation is similar to monetary inflation".
There is a degree of truth in these remarks, but much of it is explicible in terms of classical analysis of class conflict but also recognising the complexity of labour. The correlation between education, productivity, and general wealth should not be denied. Additional complexity also means additional specialist skills which narrows the traditional monopsonistic labour markets. There is, as mentioned, significant externalilities derived from population with higher education levels. Thus, whilst the insights of a conflict theory and an opportunity trap are recognised as providing insight, they are not entirely accepted in terms of substance. To the educator, this surely brings some relief - it is still useful on a social level to provide advanced skills.
Such an analysis takes issue with Brown's claim that: "If in the future everyone had a PhD or the like, then these advanced degrees would be worth no more than a job in a fast food restaurant." If we engage in such a though experiment it would be quickly realised that a society where everyone is a post-doctoral (or equivalent) graduate would not only be a pleasant society to be in, but it would also have enormous increases in overall productivity through the implementation of (a) new technologies that produce more for less and consumer goods that save time and energy, (b) new social organisation and legal norms with greater moral grounding and (c) aesthetic products that provide greater inspiration.
Another criticism of the knowledge economy approaches comes from Peter Robinson ('The Tyranny of League Tables: international comparisons of educational attainment and economic performance', in Robin Alexander, Patricia Broadbent, David Phillips (eds), Learning from Comparing Volume I : Contexts, Classrooms, and Outcomes, Symposium Books, 1999). This empirical study takes issue with the assumed correlation between educational attainment and economic performance. There is recognition that a high level enrolment of children in primary secondary education as a "necessary condition for poorer countries to reach a state of economic growtth which will allow them to converge with the living standards of the advanced industrial economies".
Beyond that level however the any supposed correlation is dubious. The evidence in advanced industrial countries linking mathematics attaintment with growth in GNP per capita is weak. There is no meaningful correlation at all with science attainment, and only a limited relationship with literacy. Indeed, once the initial eductional standards are reached there seems to be greater correlation with educational attainment and economic success where there is a higher level of income equality. In order to succeed the knowledge economy must be a distributed economy.
A significant portion of the argument from a conflict theory perspective is that the driver for learners in the knoweldge economy is individualistic and acquisitive, as opposed to cooperative and inquisitive. Indeed, it is almost a precondition for their argument. But this runs contrary to the nature to a great deal of advanced education, but in terms of the content and in the psychology of the learners. It is particularly pertinent in the field of research computing which depends very heavily on the provision and use of free and open source software. Doubtless most people in advanced research computing are hopeful that their skills shall provide a good income, but the very complexity of the work itself tends to be the cause of that income rather than some form of carefully guarded trade-secrets.
The two dichotomies however seem a little simplistic, as they overlook the possibility of not just a continuum, but rather a a multiplicity of simultaneous motivations that can occur in an individual educational decision. A person may, for example, be both sincerely interested in bioinformatics as well as motivated by the possibility of acquiring a stiff piece of carboard which allows them to use the letters "PhD" as a post-nominal. If these dichotomies are not a continuum, then what are they? Are they completely independent metrics or they dependencies? An initial consideration leans towards the latter, based merely on the observation of those who seem unduly influenced by external influence, validation, and acquisition.
It is anecdotal evidence, but it the recollection of one group of summer research students at VPAC who displayed little interest in research computing as such, but had been encouraged by their parents to seek the path of a well-paying job that generates this consideration. They had adopted, or had been influenced, to take a path that provided material satisfaction without internal joy, and as such displayed the dissonance that results in taking up acquisitive goals without an intrinsic justification. Encountering such individuals is challenging for the eductor; effort needs to be made to generate the spark of enthusiasm and love of the subject matter but if the learners have on acquisitive goals, then it is difficult to see how they will provide the motivation necessary for advanced research.
[Review for constant reference to advanced learners and computer education]
Along the continuum, adult learners are different to child learners. The educator must be aware of their particular mental and social characteristics and adjust lesson plans, teaching style, and course content accordingly. The discipline of development psychology is extremely important in making the necessary adjustments. In particular one can derive importance from Sigmund Freud on psychosexual development, Jean Piaget in cognitive development, Lawrence Kohlberg on moral development, and George Herbert Mead on symbolic development. Following the neurological changes that correlate with the onset of puberty, a person gains the capacity to engage in formal operations on a cognitive level, post-conventional thinking in moral reasoning, recognition of the other in terms of the generalised other, and the adoption of roles.
These adult developments need to be placed in the context of the sequential paths in childhood development. In the field of cognitive structures, there is a progression from sensorimotor cognition, to the preoperational, concrete operational, to formal operations from adolescence onwards. Likewise in moral development, from again the sensorimotor, to preconventional, conventional, to post-conventional moral reasoning, in psychosexual development from instinct, to the id, ego, and superego, and in interactions from the instinctual to play, games, and the adoption of roles. Beyond this however there is the period of young adulthood where the capacity for such mental structures becomes a reality. Having the mental structures does not automatically imply the knowledge or experience to put this ability into practise. Typcially stabilisation of brain plasticity from myelination of the prefrontox cortex and the pruning of synatic clusters does not end, on average, until the mid-twenties.
The distinction and continuum between child learners and adult learners is elaborated in the differences between pedagogy and andragogy. The former term is sufficiently well known, however the latter is reltively more recent. Although coined by German educator Alexander Kapp in 1833, who noted that the teaching of adults was different to the teaching of childen. The term fell into disuse until an appearance in the 1920s by Eugen Rosenstock-Huessy again arguing for special requirements in adult education. Even after that it was not receiving much consideration in the English speaking world, even if it was being used in France, Holland, and Yugoslavia (Nottingham Andragogy Group (1983) Towards a Developmental Theory of Andragogy, Nottingham: University of Nottingham Department of Adult Education). It was not until the work of Malcolm Knowles (Knowles, M. S. (1968). Andragogy, not pedagogy. Adult Leadership, 16(10), 350-352, 386.)
There are several characteristics of adult learners that an educator must be aware of. Firstly, there is a principle of adult learners have different needs for learning. Whereas younger learners may find themselves in an educational institution for statutory reasons, the adult learner a wider range of reasons. Some have particular employment or non-vocational needs to attend, where the need is externally generated - rather like the younger learner. Older learners tend, however, to have more internally generated motivations, such as desiring to improve their skills, for their own ends. There is of course, always some overlap here. Few adult learners engage in learning entirely due to internal motivations alone, least of all due with considerations of socialisation. But at least the recognition of different needs is appropriate. As related eleboration is the difference in the lhe learner's self-concept. A child learner is typically well aware if their dependencies, as much as they may disapprove of them. The adult learner is likewise aware of the responsibilities that comes with independence.
These conceptual and motivational differences correlate with differences in the learner's experience. As should be evident adult learners have a rich resource of experiences to draw upon as examples, almost invariably far greater that those of child learners. Adult learners likewise typically have taken these experiences and built existing frameworks of reality in which these experiences have been integrated. This can be, of course, positive or negative. An experienced learner may have deeply ingrained prejudices rather than deeply considered convictions as a result, and they can be more difficult to shift, especially in consideration of the relative equality of power in the learning environment between adult learners and their educators - it should not be unexpected that in some circumstances adult educators will be younger and less experienced that they people whom they are providing education to. An approach that uses the pedagogical educational authority is unlikely to have much resonance in such an environment.
As a result, the content must also differ.