On the importance of place

Distance learners and teachers in different kinds of spaceI had the great pleasure of being invited to the Open University of the Netherlands and, later in the day, to EdLab, Maastricht University a few weeks ago, giving a slightly different talk in each place based on some of the main themes in my most recent book, How Education Works. Although I adapted my slides a little for each audience, with different titles and a few different slides adjusted to the contexts, I could probably have used either presentation interchangeably. In fact, I could as easily have used the slides from my SITE keynote on which both were quite closely based (which is why I am not sharing them here). As well as most of the same slides, I used some of the same words, many of the same examples, and several of the same anecdotes. For the most part, this was essentially the same presentation given twice. Except, of course, it really, really wasn’t. In fact, the two events could barely have been more different, and what everyone (including me) learned was significantly different in each session.

This is highly self-referential. One of the big points of the book is that it only ever makes sense to consider the entire orchestration, including the roles that learners play in making sense of it all the many components of the assembly, designed for the purpose and otherwise. The slides, structure, and content did provide the theme and a certain amount of hardness, but what we (collectively) did with them led to two very different learning experiences. They shared some components and purposes, just as a car, a truck, and a bicycle share some of the same components and purposes, but the assemblies and orchestrations were quite different, leading to very different outcomes. Some of the variation was planned in advance, including an hour of conversation at the end of each presentation and a structure that encouraged dialogue at various points along the way: these were as much workshops as presentations. However, much of the variance occurred not due to any planning but because of the locations themselves. One of the rooms was a well-appointed conventional lecture theatre, the other an airy space with grouped tables, and with huge windows looking out on a busy and attractive campus. In the lecture theatre I essentially gave a lecture: the interactive parts were very much staged, and I had to devise ways to make them work. In the airy room, I had a conversation and had to devise ways to maintain some structure to the process, that was delightfully disrupted by the occasional passing road train and the very tangible lives of others going on outside, as well as an innately more intimate and conversational atmosphere enabled (not entailed) by the layout. Other parts of the context mattered too: the time of day, the temperature, the different needs and interests of the audience, the fact that one occurred in the midst of planning for a major annual event, and so on. All of this had a big effect on how I and others behaved, and on what and how people learned. From one perspective, in both talks, I was sculpting the available affordances and constraints to achieve my intended ends but, from another equally valid point of view, I was being sculpted by them. The creators and maintainers of the rooms and I were teaching partners, coparticipants in the learning process. Pedagogically, and despite the various things I did to assemble the missing parts in each, they were significantly different learning technologies.

The complexity of distance teaching

Train journeys are great contexts for uninterrupted reflection (trains teach too) so, sitting on the train on my journey back the next day, I began to reflect on what all of this means for my usual teaching practice, and made some notes on which this post is based (notebooks teach, too).  I am a distance educator by trade and, as a rule, with exceptions for work-based learning, practicums, co-ops, placements, and a few other limited contexts, distance educators rarely even acknowledge that students occupy a physical space, let alone do we adapt to it. We might sometimes encourage students to use things in their environments as part of a learning activity, but we rarely change our teaching on the fly as a result of the differences between those environments. As I have previously observed, the problem is exacerbated by the illusion that online systems are environments (in the sense of being providers of the context in which we learn) and that we believe we can observe what happens in them. They are not, and we cannot. They are parts of the learners’ own environments, and all we can (ethically) observe are interactions with our designed systems, not the behaviour of the learners within the spaces that they occupy. It is as hard for students to understand our context as it is for us to understand theirs, and that matters too. It makes it trickier to model ways of thinking and approaches to problem solving, for example, if the teacher occupies a different context.

This matters little for some of the harder elements of the teaching process. Information provision, resource design, planning, and at least some forms of assessment and feedback are at least as easy to do at a distance as not. We can certainly do those and make a point of doing them well, thereby providing a little counterbalance. However, facilitation, role modelling, guidance, supporting motivation, fostering networks, monitoring of learning, responsive adaptation, and many other significant teaching roles are more complex to perform because of how little is known about learning activities within an environment. As Peter Goodyear has put it, matter matters. The more that the designated teacher can understand that, the more effective they can be in helping learners to succeed.

Because we are not so able to adapt our teaching to the context, distance learning (more accurately, distance teaching) mostly works because students are the most important teachers, and the pedagogies they add to the raw materials we provide do most of the heavy lifting. Given some shared resources and guided interactions, they are the ones who perform most of the kinds of orchestration and assembly that I added to my two talks in the Netherlands; they are the ones who both adapt and adapt to their spaces for learning. Those better able to do this in the first place tend to do better in the long run, regardless of subject interest or innate ability. This is reflected in the results. In my faculty and on average, more than 95% of our graduate students – who have already proven themselves to be successful learners and so are better able to teach themselves – succeed on any given course, in the sense of reaching the end and achieving a passing grade.  70% of our undergraduates, on the other hand, are the first in their family to take a degree. Many have taken years or even decades out of formal education, and many had poor experiences in school. On average, therefore, they typically have fewer skills in teaching themselves in an academic context (which is a big thing to learn about in and of itself) and we are not able to adapt our teaching to what we cannot perceive, so we are of little assistance either. Without the shared physical context, we can only guess and anticipate when and where they might be learning, and we seldom have the faintest idea how it occurs, save through sparse digital signals that they leave in discussion forums or submitted assignments, or coarse statistics based on web page views. In a few undergraduate core courses within my faculty it is therefore no surprise that the success rates are less than 30%, and (on average) only about half of all our students are successful, with rates that improve dramatically in more senior level courses. The vast majority of those who get to the end pass. Most who don’t succeed drop out. It doesn’t take many core courses with success rates of 30% to eliminate nearly 95% of students by the end of a program.

Teaching with a context

We can better deal with this if we let go of the illusion that we can be in control and, at the same time, find better ways to stay close: to make the learning process including the environment in which it occurs, as visible as possible. It is emphatically not about capturing digital traces and using analytics to reveal patterns. Though such techniques can have a place in helping to build a picture of how learners are responding to our deliberate acts of teaching, they are not even close to a solution for understanding learners in context. Most learning analytics and adaptive systems are McNamara Machines, blind to most of what matters.  There’s a huge risk that we start by measuring the easily measurable then wind up not just ignoring but implicitly denying that the things we cannot measure are important. Yes, it might help us to help students who are going to get to the end anyway to get better grades, but it tells us very little about (for instance) how they are learning, what obstacles they face, or how we could help them orchestrate their learning in the contexts in which they live.  Could generative AI help with that? I think it might. In conversation, an AI agent could ask leading questions, could recommend things to do with the space, could aggregate and report back on how and where students seem to be learning. Unlike traditional adaptive systems, generative AI can play an active discovery role and make broader connections that have not been scripted. However, this is not and should not be a substitute for an actual teacher: rather, it should mediate between humans, amplifying and feeding back, not guiding or informing.

For the most part, though, I think the trick is to use pedagogical designs that are made to support flexibility, that encourage learners to connect with the spaces live and people they share them with, that support them in understanding the impact of the environments they are in, and, as much as possible, to incorporate conduits that make it likely that participants will share information about their contexts and what they are doing in them, such as through reflective learning diaries, shared videos or audio, or introductory discussions intended to elicit that information. A good trick that I’ve used in the past, for example, is to ask students to send virtual postcards showing where they are and what they have been doing (nowadays a microblog post might serve a similar role). Similarly, it can be useful to start discussions that seek ideas about how to configure time and space for learning, sharing problems and solutions from the students themselves. Modelling behaviours can help: in my own communications, I try to reveal things about where I am and what I have been doing that provide some context and background story, especially when it relates to how I am changing as a result of our shared endeavours. Building social interaction opportunities into every inhabited virtual space would help a lot, making it more likely that students will share more of what they are doing and increasing awareness of both the presence and the non-presence (the difference in context) of others. Learning management systems are almost universally utter rubbish for that, typically relegating interactions to controlled areas of course sites and encouraging instrumental and ephemeral discussions that largely ignore context. We need more, more pervasively, and we need better.

None of this will replicate the rich, shared environments of in-person learning, and that is not the point. This is about acknowledging the differences in online and distance learning and building different orchestrations around them. On the whole, the independence of distance students is an extremely good thing, with great motivational benefits, not to mention convenience, much lower environmental harm, exploitable diversity, and many other valuable features that are hard to reproduce in person. When it works, it works very well. We just need to make it work better for those for whom that is not enough. To do that, we need to understand the whole assembly, not just the pieces we provide.

The problematic metaphor of the environment in online learning (update: found a publisher!)

This is a preprint draft of a paper that has been translated by the exceptionally talented Junhong Xiao (he always gives the best and fastest feedback I’ve ever received on any of my work, and he does the translations) for publication in a forthcoming (likely August) edition of the open Journal of Distance Education in China. I’ll be touting it for publication in English so, if you’ve got an open journal that might want it or something like it, drop me a line: it’s a 10,000 word paper but I could shrink it to fit journal needs if that’s too long (thanks to editors of the OTESSA Journal for taking this on!). The paper is in fact mostly a mashup of a couple of two of my earlier blog posts – Nobody has ever learned anything at a distance, and no one ever goes to a distance institution  and A few thoughts on learning management systems, and on integrated learning environments and their implementation though it comes to some slightly different conclusions and emphasizes a few different things (and it has more references!).

I was reminded to share this because I attended an excellent and thought-provoking opening keynote yesterday by Martin Weller at the OTESSA 2022 conference, in which he discussed themes and ideas from his forthcoming AU Press book, Metaphors of Ed Tech. Martin takes a much broader (and really interesting) perspective on uses of metaphor than I use in this paper: I’m really looking forward to reading the book. This paper is largely focused on some of the more obvious spatial metaphors, notably that of the ‘environment’. I’m releasing it as CC-BY-NC so do as you wish with it but, if you do, please give credit both to me and to the Journal of Distance Education in China, where it will be published in Chinese (trans. Junhong Xiao).  Sorry for any weirdness caused by copy-and-paste from the original.

The problematic metaphor of the environment in online learning

Jon Dron, Athabasca University, jond@athabascau.ca

Abstract

In online educational systems, teachers often replicate pedagogical methods, and educational institutions replicate systems and structures used by their in-person counterparts, the only purpose of which was to solve problems created by having to teach in a physical environment. At the same time, a great deal of the development and use of learning technologies has focused on creating virtual learning environments that attempt to replicate features of their physical counterparts, thereby weakly replicating in software the problems that in-person teachers had to solve. This has led to a vicious circle of problem creation and problem solving that benefits no one. In this paper I argue that the term ‘environment’ is a dangerously misleading metaphor for the online systems we build to support learning, that leads to poor pedagogical choices and weak digital solutions. I propose an alternative metaphor of infrastructure and services that can enable more flexible, more learner-driven, and more digitally native ways of designing systems (including the tools, pedagogies, and structures) to support learning.

Keywords: online learning, learning environment, learning management system (LMS), Next Generation Digital Learning Environment (NGDLE), personal learning environment (PLE), learning infrastructure.

Introduction

Outside the walls of educational institutions, for those with adequate Internet access, intentional learning using online systems is almost certainly more popular than its in-person counterpart, as at least the first port of call for learning almost anything and, often, as the primary means through which it occurs. From Google Search to Wikipedia, from MOOCs to Twitter exchanges, from YouTube videos to Khan Academy tutorials, people with online access are swamped with  learning opportunities. However, many academics and students still see online education as a poor second-best to in-person learning (e.g. Protopsaltis & Baum, 2019; Bouygues, 2019; Tichavsky, Hunt, Driscoll, & Jicha, 2015). In this paper I will argue that the distinction between online and in-person learning is far less significant than it appears, because all learning is in-person and never online, and most learning that is labelled as ‘in-person’ actually occurs at a distance from the teacher. Problems emerge, however, when  institutional online teaching inadequately attempts to replicate features and forms of in-person teaching, many of which:

  1. Exist to solve problems caused by the distinctive physical, temporal, psychological, and economic limitations of material spaces, and
  2. Are successful mainly as a result of the features and forms of the physical setting, not as a result of intentional teaching.

As a result, the systems we develop may not take full advantage of the medium, may not take advantage of the physical context of the students, and may attempt to solve problems that should not exist for those using them to learn because they are the result of in-person constraints. In many cases, online teaching may therefore actively militate against effective learning. Some of the problems may be solved using pedagogical adjustments in teaching and organizational changes at an institutional level, discussed in the first part of this paper. However, many emerge from the electronic systems that we use to teach, that poorly mimic the functions of their in-person counterparts in software. In this paper, I suggest that this is, to a significant extent, due to the misappropriation of spatial metaphors that cannot and should not be applied to online systems. I propose a different approach to the construction and conceptualization of tools for online learning, that better reflects the innate benefits of the medium, and that more fully supports the needs and circumstances of both online students and online teachers. I conclude by putting the pieces together and suggesting ways that, in combination, pedagogical, organizational, and digital changes may co-evolve to achieve the potential transformation of education that is afforded by digital networked devices.

In-person teaching

Although they may have physical or virtual windows to the world outside, the walls of the classroom provide clear boundaries that define the space and, because participants must be co-present,  the time in which activities intended to bring about learning occur. This is also true of most of the other buildings, rooms, and spaces that are provided by in-person institutions for students, including corridors, student accommodation, meeting rooms, common rooms, cafes, halls, quadrangles, staff offices, libraries, gymnasia, and examination halls. These spaces are not just support structures for the classroom, but active participants in the learning process (Dron. 2021). Even the act of physically walking to the classroom, especially with other people, creates a salience and value to the activity that is very different from that of clicking a link to an online resource. Most significantly, they are social spaces where learning happens as a result of direct and indirect interactions between learners (who are one another’s teachers) and, often, with their designated teachers. Simply seeing others learning makes a difference, as do the fliers and leaflets on the walls, and the spaces intended to support clubs and societies, where much academic discussion often occurs. Cafes, bars, and canteens are rich in learning dialogue. Student rooms, dormitories, and (especially) their kitchens, are powerful seed beds for learning, where much sense-making discussion occurs. Many universities provide purpose-built study areas. Even and perhaps especially, areas cordoned off for smokers provide an extremely fertile space where students from different subject areas and disciplines can and do share and construct their knowledge. Similarities in their design the world over speak to the fact that these are highly evolved spaces, supporting a learning process that extends far beyond the classroom.

These countless diverse learning opportunities in physical spaces, perhaps counter-intuitively, speak to the fact that there is a distance component to virtually all in-person education. Indeed, almost all learning is distance learning, in the sense of occurring somewhere and somewhen other than where and/or when deliberate, instrumental teaching occurred. People who learn with teachers in a physical space are almost always also interacting with other participants in the teaching role at a distance, usually in time and space, such as textbook authors, classroom designers, editors, illustrators, creators of timetables, and curriculum designers. And, for ‘in-person’ institutional learners, much of the learning itself also occurs at a distance, outside the classroom. This is most obvious in the form of assignments and homework but just as much learning can occur in conversation and interaction with others. Even when alone, if teaching works, sense-making connections always occur after the lesson is over, and continue to do so long after (sometimes decades after) the teaching event, almost never in the same place that the lesson originally occurred. In-person students do not have one teacher: they may have thousands. Weaknesses in in-person teachers can often be compensated for by these many other teachers, including the learners themselves and the institutions that provide the framework and resources for learning. This is amplified by that fact that, although credentials and grades are highly antagonistic to persistent intrinsic motivation (Kohn, 2011; Blum & Kohn, 2020; Ryan & Deci, 2017), they do encourage compliance. In search of good grades, students will therefore make use of whatever means they can – including those many other teachers as well as cheating or satisficing – to achieve the marks they seek. The physical environment of an in-person institution provides many supports to make this possible.

Online teaching

The in-person teacher, by default, controls learning in the classroom because it is a self-contained environment of which they are in charge for the duration of the lesson. Relinquishing control must be an active choice, or the result of an error. In contrast, the online teacher cannot, without a great deal of concerted effort, control the online student, any more than a writer of a book can control a reader. Online students can always choose when, where, with whom, how, for how long, and with what tools, media, and resources they learn (Dron & Anderson, 2014). It would therefore be surprising were online pedagogies to closely resemble their in-person counterparts, because they have different problems to solve. Most notably, without the requirement to share a single environment, with all the many rules, norms, structures, and constraints that entails, and without the need for the teacher to fill every moment of classroom time with learning activities, there should be no need for teachers to exercise the same level of control over their students.  However, online teaching evolved from in-person teaching, and online institutions must continue to interoperate with the in-person educational systems of which they are a part. As a result, many online teachers assume that they should dictate the learning process as much as their in-person counterparts and, usually, it becomes a partly self-fulfilling assumption through coercive methods like frequent grading, draconian scheduling, and tests. They consequently often make use of very similar pedagogies to those of their in-person colleagues, struggling to find simulacra or workarounds for the affordances of physical spaces that are no longer available, vainly believing that the learner is going to follow the path that they have determined for them and, too often, imagining that this is the sum total of the learning experience. To make matters worse, educational institutions impose other structures that are purely the result of constraints of teaching in physical classrooms, such as fixed-length (or multiples of fixed lengths) courses, deadlines, and perhaps most perniciously, the concept of failure. As any game-player or musician knows, failure is part of learning: it cannot ever be its end but, because of the constraints of having to run a course with co-present students and a beginning and end, failure becomes a potential outcome, not just part of the process. When all these factors are put together, the online student may have little more independence than their in-person counterpart but, at the same time, may lack the countless structures and forms of physical institutions that support in-person students. Rather than being immersed in learning opportunities, they must actively seek them within their own physical and virtual environments.

One obvious solution to this problem would be to create an online learning system that provides much of what is lost in translation from the physical environment, for example through a custom-made social media platform or an informal discussion area within a learning management system. However, this is not as easy or effective as it may seem, especially if it remains tightly coupled with other institutional policies, norms, and teaching methods. Partly, this is because of the too common focus on explicit outcomes and grading found in most institutional teaching together with failure by students and teachers to recognize the critical role of in-between spaces in learning. Thanks to the extrinsic coercion of marks and credentials,  if it makes no direct contribution to a grade, then it is seen as less valuable. Mainly, though, it is because it is not just there: students will not pass it on their way to somewhere else or be there for other reasons (like a need for rest or refreshment). They have to intentionally visit, typically with a purpose in mind. However, as the main value of it is its purposelessness (or, at least, that it supports a very broad set of purposes), that is rarely going to happen. Online systems are not environments in which students dwell: they are parts of their own environments.

This speaks to the central phenomenon around which this paper revolves: that nobody actually learns anything at a distance. We are always learning it where we are now. All learning is in-person learning, and it all takes place within a physical environment, part of which (but only a part) may include whatever technologies we might be using to talk with people, read, watch, listen, and learn from: books, computers, pens, emails, learning management systems (LMSs) and so on. Some of these may extend into other physical spaces occupied by other people, perhaps at other times, connected by online means. The broader learning environment is highly distributed in time and space, but learning itself only occurs locally. What we describe as ‘distance learning’ or ‘online learning’ is thus, in fact, nothing of the kind. It may involve distance or online teaching but the learning is always in-person. Online students exist, because the word ‘student’ only has meaning in relation to an online teacher, but online learners do not.

The promise of online learning environments

It is understandable that, when we teach in person, we have to occupy and make different uses of the same or similar environments like classrooms, labs, workshops, lecture theatres, and offices. There are huge financial, physical, and organizational constraints on making the environment fit the task, so it would normally be madness to build or even to substantially reconfigure a whole new classroom every time we wished to run a different class. Rooms may be built for flexibility, with moveable partitions and furniture, and that is much to be wished for, but there are physical limits such as walls and property boundaries that prevent this from going too far. Instead, our pedagogies and processes are normally made to fit the affordances and constraints of the classroom: they are another problem that our pedagogies have to solve, and/or an opportunity that our pedagogies can take advantage of. We may, sometimes, have some choice between classrooms that offer different facilities but, for the most part, our options are limited by what has already been built.

Online, there are countless tools available and, if none are suitable, it is not too hard to build them or to modify them to suit our needs, at least when compared with the costs of creating new physical spaces. There are few significant physical limits on how many can be used or how many people may use them: there are none of the limitations of physical space that constrain the use of buildings. Once they are built, moving between virtual tools just takes a tap of a screen or the click of a mouse or keyboard. It is even possible to use several of them at once, especially with a large high resolution monitor or more than one device. The learner’s environment may contain countless tools and systems, any of which may support learning, including physical books, instruments, and other people around them. And yet, for the most part, online teachers tend to make use of only a handful of possible tools: most consist of no more than a learning management system , email, and perhaps a webinar system.

There are many mutually reinforcing reasons that online teachers rarely provide the perfect application or combination of applications for the context of study:

  • Teachers’ lack of knowledge of the options (it takes time and effort to discover what’s available).
  • Teachers’ lack of skill in using them (most interesting tools have a learning curve, and that gets steeper in inverse proportion to the softness and diversity of the toolset, so most teachers don’t even know how to make the most of what they already have).
  • Lack of time and/or money for development (an application is just a shell for the content it contains and the interactions it supports, and it is not always as easy to add existing materials to a new tool as it might be in a physical space: for example, an in-person lecturer only needs to talk, whereas an online teacher must master the complexities of the hardware and software needed to record, edit, and share the same thing ).
  • Costs and difficulties in management (each tool adds costs in managing faults, configuration, accounting for use, performance, and security).
  • Cognitive load involved for learners in adapting to the metaphors, signposts, and methods needed to use the tool itself.

All of these are a direct consequence of the very diversity that would make us want to use different applications in the first place. This is a classic Faustian Bargain (Postman, 1998) in which the technology does what we want, and in the process creates new problems to solve.  Every digital system must establish rules of engagement that its users must learn, such as the ways that navigation occurs, the ways to make it perform its functions, the terminology it uses, and so on. In effect, every application invents its own metaphorical physics. That makes virtual systems harder to find out about, harder to learn, harder to develop, costlier to manage, and more difficult to navigate than the static, fixed facilities found in particular physical locations. They are all different, there are few if any universals, and any universal today may become a conditional tomorrow. In the case of cloud-hosted systems, the owners of which may unilaterally make changes to the software or configuration, this may be literally so.

Learning management systems

The learning management system (LMS) addresses all of these problems, to some extent. Almost every LMS essentially automates the functions, though not exactly the form, of traditional classrooms. Indeed, they are typically seen as environments, or are referred to as ‘platforms’, underlining the physical metaphors that inform them. In some parts of the world people prefer to use the term ‘managed learning environment’ (MLE), and the LMS/MLE is, in most vocabularies, the most dominant representative of a larger category of systems usually described as virtual learning environments (VLEs) that also includes things like MOOs (multi-user dungeons, object oriented), immersive learning environments, and simpler web-based teaching systems that replicate aspects of physical teaching such as Google Classroom or Microsoft Classroom. The use of spatial metaphors for the names of such systems reflects a deep-held belief or tacit assumption that the virtual systems can provide the boundaries within which actions occur, in ways that tend to be seen as analogous to those of physical spaces. In a few limited contexts, notably through immersive systems, this belief may be partly justified,. However, it matters that even the most immersive system occurs in a physical space. For instance:

  • when participants leave the immersive system they exit into different rooms, losing the natural opportunities for incidental or continuing chat that are innate to physical spaces;
  • participants are at the mercy of dropped network connections, glitches, and issues with the machines that run the immersive environment, leading to potentially quite different experiences for different participants;
  • different participants experience different temperatures, background sounds, smells, and opportunities for interruption in their own physical environments.

LMSs differ from physical environments to a much greater extent than immersive systems, in ways I will describe over the rest of this section. This is not a trivial issue of nomenclature. I will be arguing that the misconception that they are meaningfully analogous to physical classrooms lies at the heart of many weaknesses and failings in both the design of the tools and their use, reinforcing the belief that online teaching closely resembles in-person teaching, and blinding us to essential differences between the two.

The building metaphor

Creators of early LMSs and VLEs back in the 1990s (including the author) based their designs on the functions and entities found in a traditional university because that was the context from which they sprang, and that was the context in which they had to fit. In the eyes of its designer, an LMS could be thought of as a big university building with rather uniform classrooms. It may have extensions built onto it using plugins or standards such as LTI (the learning tools interoperability standard), and it may have a few doors and gateways (mainly in the form of hyperlinks) linking it circuitously or in jury-rigged fashion to other similarly weakly connected ‘buildings’  such as ‘places’ to register, to seek support, to talk to an advisor, to complain, to find books, and so on. For the most part, though, its fundamental organizational metaphor is that of a university, college, or school.

 

The LMS is, however, an impoverished school. It has no metaphorical corridors, halls, common rooms, canteens, yards, libraries or any of the other parts of a typical university environment where students gather to (amongst other things) learn. Students rarely get to even be aware of other classrooms beyond those they are in. Some teachers may give classrooms informal-sounding names like ‘the learning cafe’ but it is still just another classroom that works in the same way as the rest. Students teleport from one classroom to the next because what happens in between is not perceived by the designers as a useful classroom function to be automated or perhaps, more charitably, they could not figure out how to automate that. But there are other differences that are, perhaps, even more pernicious, to which we turn next.

Centralized code bases

In a physical environment, every object is discrete, occupying its own space. Physical classrooms can and do change – new furnishings, equipment, and so on – but the effects are local to that particular space, and they seldom prevent teaching from occurring across an institution.  Learning management systems, on the other hand, re-use the exact same code to generate all of the virtual classrooms of which they consist. Instead of a number of courses occupying the same physical spaces, and there being many such spaces to choose from, every course gets its own instantiation of a single centrally hosted toolset. There may be options to switch features on and off within any given course instantiation, options to configure each component differently, and a choice between components may be offered, but everyone gets exactly the same set of features, determined by the developer and the system administrator. This means that one set of features has to suit everyone. If, say, a teacher wants a discussion component that does things the default discussion component does not support, then it has to be installed or integrated in the centralized code base. While the LMS may technically support this – through plugins, LTI integrations, OKI components, and so on – system administrators are usually rightly reluctant or unable to allow it. Every component is another potential source of failure or (often) security holes, incurs management costs, uses system resources, creates a significant maintenance burden, and increases the complexity of the system for everyone. To allow unfettered installation of alternative components would be completely unmanageable. As a result, most available features must be a compromise, that can be bent to suit the needs of (typically) thousands of courses and teachers, but that are unlikely to be an ideal fit with any of them. Unlike the physical classroom, changes to the underlying application affect everyone, at once. When the LMS goes down, it takes the whole institution with it, and when changes are made, they are made for everyone, often affecting hundreds or thousands of courses and tens of thousands of students.

This is particularly problematic in cloud-based systems where administrators are not even part of the same organization, and where the system must support hundreds or thousands of institutions. Few of us who teach using cloud-based systems have not experienced difficulties when the systems on which our courses run change without warning or consultation, disabling or altering things that disrupt the design, sometimes rendering it inoperable. Even when they work, the fact that they use a single code base limits the potential for customization. Because most LMSs based their designs on what was presumed to occur in an average university, they rarely fit well with any actual university, because virtually no universities are average. Sometimes, the problems may be relatively minor. For example, Blackboard calls its organization elements  ‘courses’, whereas many other names for such things are common, including modules, units, and papers, and ‘course’ may refer to what others around the world might call a ’program’. Even this may disrupt and cause confusion (Dron, 2006). Other problems can run deeper, to which we turn next.

Reified roles

The typical LMS is a very controlled environment where everyone has a programmatically enforced role (typically at least partially reflecting traditional educational roles), that may vary according to the ‘room’ in question, but that are far less fluid than those in physical spaces. There are strong hierarchies, and limited opportunities for moving between them. Some of those hierarchies are native to the online learning system: the system administrator, for instance, has far more power than anyone in a physical university to determine how learning happens, like an architect with the power to move walls, change the decor, add extensions, and so on, at will. The programmers of the system are almost god-like in their command of its metaphorical physics. But the ways that they give teachers (or learning designers, or administrators) control, as designers, directors, and regulators of the classroom, are perhaps the most pernicious. In a classroom a teacher may lead, and that is the default, but they may and usually should choose to at least share leadership with their students, often fluidly and in response to how students are learning. In an LMS, a teacher (or someone playing that role) must lead,

Tools such as discussion forums may seem to be more egalitarian, but teachers’ power to control events in them is usually far greater than that of their in-person colleagues, often including the means to delete unwanted messages, prevent replies, stop conversation threads stone dead, and many other things that would be superhuman capabilities in a physical space. In a physical classroom, a determined enough student can always make themselves heard. In an LMS, the teacher can silence them. There is thus less of the soft flexibility found within in-person classrooms that allows for conversational pedagogies that adapt to the interests and needs of learners. At the same time, though, it should be (though too rarely is) remembered that the teacher’s power is confined to a small part of the learner’s own environment, not to a whole classroom. In practice, teachers still tend to treat the forum as an analogue of the classroom and, recognizing the value of dialogue in such contexts, often resort to coercion to make it happen online: marks for discussion contributions are far more common than in in-person settings, even among experienced online teachers. This combination of hard, role-based digital authority and hard, reward-based pedagogical authority is fundamentally different from its physical analogue. It creates both a social and a power distance that compounds what is already a less immediate relationship between student and teacher.

Within the LMS the teacher sees things that students cannot, and controls things that the students may not. A teacher configures the space, and determines with some precision how it will be used. With a lot of effort and (usually) high risk to the security and stability of the system, it can be made to behave differently, but it almost never is, because doing so usually involves promoting students to roles with similar capabilities to that of the teacher. In many cases, especially when it involves the use of plugins or other tools that extend across the system, this cannot be localized, so the risks to every user of the system must be considered. This is beyond the capabilities or rights of most teachers, and so it usually falls to system administrators, reinforcing their already substantial power to affect the teaching process.

Functional design

An LMS is typically built along functional lines. Rather than attempting to be a precise mirror of the in-person context, its functions are mostly based on loose, superficial observations of the things that teachers and students seem to do in physical classrooms, analysed to their component parts. Mostly, they are structured by teaching functions: presenting, discussing, assessing, guiding, and so on. For instance, in most LMSs, if you want to talk with someone, you normally need to go to a separate discussion area inside the classroom or, metaphorically,  to leave a note on the teacher’s desk in the form of a direct message. Unlike a physical classroom, dialogue is seldom possible everywhere. The same is true if you want to take a test, or to share your work with others: it rarely occurs within the context of learning, but in a separate screen, often separated from its context by a hierarchical set of links. Indeed, in many architectures, it will be handled by a different component than the rest, with its own tables in the database and its own distinctive interface.

Similarly, lectures are either literally that (video recordings of lectures) or (more usefully, from a learning perspective), text and images to be read on screen. This results from the erroneous assumption that the only function of lectures is information transmission, which is perhaps their least useful role, given that we have known for almost a century that it is far more effective to read a book (Greene, 1928). Lectures can and do have value as physical and temporal signposts, as motivators to pay attention, as events that demand attendance and thus have greater salience than simple reading, as well as providing opportunities to engage with others, sometimes within but always outside the lecture hall. Online, there is seldom a chance for students to even put up a metaphorical hand to question the teacher, and ‘joining’ a lecture is no more salient than clicking a link to a Facebook post. There are limited opportunities to be aware of what other students are doing, including for the teacher (although teachers do usually have access to system logs that offer an impoverished caricature of what students are doing, albeit one that is blind to anything they do beyond clicking and tapping keys on a machine). Much of the ‘space’ may as well be unpopulated, given the little students see of one another. Learning resources are normally static and designed in advance, and so the teacher cannot nimbly adjust to student reactions to them. Notices can usually only be pinned on the ‘wall’ by teachers, often with names such as ‘announcements’, further emphasizing the controlling nature of the teacher-student relationship. Classroom timetables are embodied in software despite the fact that a rigid and unforgiving timetable makes little sense in a medium that supports learning anywhere, any time. Some LMSs may allow you to break up the content differently, but it is still another timetable; just a timetable without dates. It is always the teacher (or one to whom the role is delegated) who sets the order, pacing and content.

Robot overlords

The LMS provides a high-tech classroom, populated by metaphorical robots.

Some of the robots may be programmed to attempt to force students to behave in ways determined by those higher in the hierarchy (sometimes teachers, sometimes administrators, sometimes the programmers of the software). For instance, adaptive systems might act as gatekeepers that prevent students from moving on to the next section of work before completing the current one, or they might prevent students from submitting work before or after a specified date (Martin, Chen, Moore, & Westine, 2020), or they might limit their access to a specified time period.

Some of the robots might even mark your work (Keuning, Jeuring, & Heeren, 2018). Human beings have grown up with other humans and therefore understand the context of the work, the motivations of the students, and the many different ways that things can go wrong, as well as creative and unexpected ways they can go right. Robots – even those that are employ deep learning and similar AI approaches – do not. While hard, mechanistic systems may be useful for providing feedback when students must play their role correctly in hard, mechanistic systems (in hard, ‘right answer’ subjects), those mechanistic skills are seldom the most important part of what they learn. Human teachers do not (or should not) just judge success or failure: they should model practice,  remedy misconceptions, provide encouragement, and so on.

There are metaphorical surveillance cameras everywhere, recording students’ every move (in very low resolution), often only accessible to those with more powerful roles, though sometimes a robot or two might give them a filtered view of it, such as through learning analytics traffic-light interfaces (Verbert, Duval, Klerkx, Govaerts, & Santos, 2013). Though the perpetrators of these tools may claim to have student interests in mind, and will often talk of ‘personalization‘ by way of justification, it is not personalization at all: it is system-enforced customization done to, not by the students (Kohn, 2015). These are all tools that are designed to enforce compliance: an attempt to embody in software the control that is demanded of an in-person teacher due to an accident of physics, not for any pedagogical purpose.

Beginnings and ends

The fundamental social form of the classroom that provides the primary metaphor of most LMSs is the formal group (Dron & Anderson, 2014). Formal groups are technological entities – inventions that are designed to address problems – at least as much as they are social. Among their many technological features are names, roles, procedures, rites of joining and leaving, rules of behaviour, schedules, beginnings and ends, almost all of which arise from the constraints of in-person learning, such as the need for people to be co-present, problems when people talk at once, limits to the capacity of classrooms, directionality of hearing and sight, and so on (Dron, 2016). Unsurprisingly, many of these features are embodied in code, not only in the reified roles already discussed but in processes of joining and processes of leaving.

A student cannot usually go back and visit when their course is over because most online courses have opening and closing enrolment dates. Perhaps their designers assumed that, when teaching was done, the learning was done which, of course, it never is. Learning keeps on evolving long after explicit teaching and testing occurred. Again, this is because physical classes are scheduled and terms come to an end because they must, not because it makes pedagogical sense. And, like almost everything, it is possible to override this default, but hardly anyone ever does, partly because it brings back those Faustian bargains, especially in manageability, but mainly because most people accept defaults (Kelly, 2009, Dron, 2006). LMSs embody enrolment technologies as much as they do teaching technologies and, in the process, they unnecessarily limit potential for learning.

Because the primary metaphor of almost all LMSs is the classroom, they can be a particularly poor fit with ways of teaching that have no classes, such as self-paced courses and MOOCs, individual projects, or flexible networked ways of learning such as those underpinned by Connectivist, or Rhizomatic models of learning. This is not to say that such uses are impossible. For example, assumptions about class schedules that are embedded in software (such as that all students must submit work by a certain deadline) can be disabled, or bypassed by setting a deadline in the far distant future, then manually informing students of when to submit their work. However, the fit with self-paced models of learning is typically poor. Among the many peculiarities that result are students who engage in discussions with ‘classmates’ who no longer have access to the provided forum, and the impossibility of collaboration when every student is at a different point in the course. More challengingly, and unlike teacher-paced courses in which the teacher can modify almost any aspect of the content or curriculum at will, knowing that the whole class will be affected in the same way, much confusion and even dismay can arise when changes are made to materials that may be in use by existing students.

Imperfect caricatures of physical spaces

In summary, most LMSs provide an automated set of metaphorical classrooms that harden many of the undesirable side-effects of educational systems in software, in ways that have little to do with how best to teach, and that inappropriately apply spatial metaphors in ways that conceal rather than illuminate their functions. Each bit of automation and each navigational decision hardens pedagogical choices, at least as much as the walls, doors, and physical limitations of physical spaces and, often, more. Programmers do not replicate physical classrooms but instead create or enlist new laws, new kinds of structure, and new kinds of hardened process that can be embodied in code. Classrooms solved problems of physics for in-person teaching and form part of a much larger structure that has evolved to teach reasonably well. LMSs just focus on a limited subset of teaching roles, and empower the teacher in ways that caricature their already excessive dominance in the classroom, that only occurred because of the nature of the physical space and the constraints it imposed.

LMSs leave much to be desired, but the metaphors on which they are based bear enough resemblance to physical reality to be readily understood by teachers and students. They usually provide just enough configurability and flexibility to more or less adequately work as teaching tools, for everyone, almost no matter what their level of digital proficiency might be. They more or less address the Faustian bargains listed earlier, albeit they normally do so by stifling what we wanted and should have been able to do in the first place with online tools, In the process they create new and quite extensive problems, as well as failing to replicate most of what makes physical universities work in the first place. Virtual learning environments are not like physical learning environments: they are only ever parts of them. There are other electronic ‘places’ to escape from them, such dedicated social media, or even just plain old email, but then all those Faustian bargains come back to haunt us again. They occupy space within the learner’s own physical environment, but it is rare for pedagogical designs to even acknowledge that, let alone to consider it in the design.

Improving the LMS

It is tempting, faced with these problems, to assume that they could be solved if only we made the LMS more closely resemble the physical environments on which it is modeled. However, this is a poor solution because, as we have already seen:

1)    physical environments create constraints and problems to solve that are unnecessary and avoidable in virtual systems; and

2)    it is not practical nor is it within our technical reach to replicate all the many incidental benefits of physical environments.

That said, there are lessons to be learned from physical spaces. Among many improvements that could be made would be:

1)    To make every part of the system at least potentially social: to allow synchronous and/or asynchronous dialogue to occur on every page or screen of the system. This is the default in all physical spaces: talking has to be prohibited if it is not wanted.

2)    To allow at least some parts of the system to be free of roles, or with more flexible roles, allowing all members of the system to create and share posts and resources using discretionary access control (so it is the poster’s responsibility to choose who can see it, and who can change it). Even in highly controlled physical environments, we choose what we reveal and to whom.

3)    To support social networking and the blurring of boundaries between areas, tools, and features of the site, so that courses are just one of many kinds of organizational unit, with selectively permeable boundaries through which others can pass, or with which they can overlap. Again, this is a default in physical spaces, that leak information through walls, floors, windows, and doors, that exhibit continuity of engagement when people enter or leave classrooms, that allow teachers to open doors to others, that admit a multiplicity of primary uses.

Though these improvements appear simple to achieve, adapting an existing mainstream LMS such as Blackboard, Canvas, Moodle, or BrightSpace to support them is fraught with difficulty.

By far the easiest of these improvements to make within an existing LMS is to make it more social. Achieving this within an existing course structure is a simple programming problem that can readily be solved in countless ways. In most LMSs, it could be built as a plugin. Existing architectures, in which courses and roles play a primary structural role, make it somewhat more difficult to extend such dialogue beyond the boundaries of the course. The metaphorical walls of a course are, for the most part, more of a barrier to engagement beyond it than those of a physical classroom because their metaphorical physics can be (and is) enforced in code. It is not, however, an insoluble problem. For example, a context-aware embeddable discussion system such as Disqus or Isso, hosted locally or remotely, could fairly easily be added.

Making the system free of roles is much more difficult because, in most LMSs, they underpin almost every function and structure of the system, and they cannot be made to work with an open, discretionary access- based model of permissions: the two approaches are, architecturally, mutually exclusive. One way of dealing with this would be to follow the lead of the Drupal content management system to support ‘organic’ groups: limited areas of the LMS where everyone has the same rights to create shared content or social areas, and where anyone can control who can see what they post. These areas could be as large or as small as desired but it would be difficult to make them extend beyond a course, or to encompass one or more courses. It would not be impossible, but to do it safely and reliably (without giving everyone a single, very powerful role) would require a major rewrite of the underlying LMS.

For all of the LMSs of which I am aware, the most difficult of all these improvements would be to blur the boundaries of the tools, features, and courses. The course is such a fundamental architectural unit of most, if not all, LMSs that changes to its operation would demand a significant redesign. It could be done, but it would not be the same kind of system any more.

It is for these reasons that, wishing to support all of these features and realizing the extreme difficulty of modifying the LMS without compromising some or all of its existing functionality, a group of us at Athabasca University created The Landing (Dron & Anderson, 2014), as a separate system to the LMS, linked only by a single sign-on and tenuous hyperlinks and, to a limited extent (only supporting public posts in either direction) RSS feeds. Further efforts to design deeper integration proved too difficult, for both technical and organizational reasons. Unfortunately, The Landing suffers from the same Faustian Bargain that besets all attempts to expand the range of systems available. The maintenance burden of a system with many thousands of users is too much to sustain for a system with very limited central support and even more limited funding. Pedagogically, the system fulfills an important need and so it has survived for more than 12 years but, technically and from a management perspective, its future is in jeopardy. Similar issues are playing out the world over. The more control and diversity that we enable, the more difficult and expensive it is to manage it.

Alternative approaches

Incremental improvements

Athabasca University is currently building an Integrated Learning Environment (ILE) that centres around very conventional elements of a institutional teaching system: an LMS, some relationship management tools, a student records system, an enrolment system, an examination management system, and so on. These are tightly integrated, but it is intended that the ILE will also embrace many other tools and systems that are far less institutionally bound, from the aforementioned Landing, to other social media (such as WordPress), to portfolio tools, to shared software repositories. This is an approach that starts with replicating existing structures and services by building a tightly managed administrative core, but that is intended to grow to support more open, diverse, and rich approaches to learning and teaching, co-evolving with methods and pedagogies that are more in keeping with the different problems and needs of distance learners. However, though it provides a managed approach to supporting change, this approach carries many risks.

A design approach that treats online systems as environments invariably makes the assumption that it is where everything associated with what goes on inside it happens, and (for online systems) this creates quite unnecessary restrictions on what can happen. Athabasca University’s design approach for its ILE was highly participative, engaging most of its teaching, technical, and administrative staff and asking for what they needed. However, inevitably, their requests were based on assumptions formed by their existing practices and, especially, by the existing environmental metaphors of the LMS and associated systems with which they were already familiar. In essence, they were asked what kinds of spaces they needed, and what kinds of stuff needed to be in those spaces for them to do what they currently do.  ‘Space’ and ‘stuff’ are what Stewart Brand (1997) describes as inevitably being the fastest-changing, most volatile parts of any physical building, after site (its physical limits), structure (what holds it up), skin (mainly the external walls), and services (electricity, gas, network wiring, etc). More abstractly, this is a solid structural principle that applies as much to ecosystems and educational systems as it does to buildings. As Brand himself observes, drawing from O’Neill, DeAngelis, Waide, & Allen (1986), the larger, slower-changing elements of any system affect the smaller, faster-changing more than vice versa. In physical spaces, these naturally tend to be bigger and/or more difficult to change, but the same is true in virtual spaces, where size seldom matters that much, but hardness (inflexibility, brittleness) has the same effect. The more difficult it is to make changes, the more an element of the system determines the behaviour of other elements in the system that interact with it. The ILE’s structure, skin and services have been designed based on needs determined by perceptions of the space and stuff within it, that were in turn very strongly determined by the LMS and other systems that went before, with all the inherited baggage that they inherited from in-person environments. Hence, the ILE’s fundamental design model is really no more than an extended LMS, and it inherits most of its weaknesses. The main way in which it differs is that it is designed to be extendable, but those extensions will still – in terms of how they are treated and used – be part of that same environment, with all the aforementioned problems that this entails.

Integrated learning infrastructures

I have argued that a better name for the system being developed at Athabasca University is not an ‘integrated learning environment’ but an ‘integrated learning infrastructure’ (ILI). In metaphorical terms, it should be like the utilities, services, and mechanisms that make an environment possible, but it should never be thought of as the environment itself.

Stripped to their essentials, digital systems intended to support the educational process provide services, consisting of tools that may be used to support learning, teaching, accreditation, and other roles and functions of an educational system.  Such services are many and various: discussions, presentations, file sharing, assignment submission & grading, quizzes, blogs, scheduling, wikis, bookmarking, real-time communications tools, enrolment systems, identity management systems, support systems, and much more. There is no good reason that these should be confined to loose approximations of their physical counterparts, nor is there any good reason that teachers or system administrators should be the only ones to control them, though it is important that each of them is owned by someone, otherwise the resulting free-for all would be difficult to manage. Microservice architectures that support such systems are quite mature, and widely implemented in different fields, if not so much within the educational sector. From the point of view of end users, these can be thought of as assemblable components, and the assemblies can be performed by anyone, including students. Ideally, it should be possible to integrate them with other applications and services offered beyond an institution, including on the desktop of individual students.

Ideally, it should be possible to assemble them into units with value in the system, that can themselves be assembled into other components. This provides a path for evolution from existing approaches because those units might include courses. There may be a need for additional services to support non-teaching functions associated with educational systems, such as administration or credentialling.

Such services are not so much environments as they are infrastructure that exists within and between the different environments that learners, teachers, administrators, and technicians occupy both virtually and in person. Non-exclusively, such infrastructure may minimally support needs such as:

  • Dialogue and interactions between participants
  • The presentation and curation of content
  • Assessment, formal and informal
  • Sharing of words, images, video, audio, and other document types
  • The formation of groups,  networks, and sets (social gatherings around shared interests or other commonalities)
  • Sharing of tools and resources
  • Etc.

What matters most is that all of these services can be combined in indefinitely many ways, by anyone.

This is not a new idea. In the early 2000s, the ELF (e-learning framework) and OKI (Open Knowledge Initiative) both attempted to provide ways to assemble services (ELF) or components (OKI) in many different ways. However, for the most part, both of these initiatives were firmly focused on building centralized systems that replicated the functions of an LMS, so they carried forward the assumption that what would be built from the components would be teaching environments; a better LMS, but still an LMS.

Around the same time as ELF and OKI were being developed, and driven by similar intents,  the notion of the personal learning environment (PLE) became popular, though with very many quite radically different interpretations (Martindale & Dowdy, 2010), ranging from institutionally controlled systems that were often described as ‘platforms’ (Yen et al, 2019) to collections of applications and services assembled by a learner on their own desktop in an ad hoc fashion (Wilson, 2008). Though some of the promoters of the concept saw the environment as extending beyond virtual systems, the vast majority of these interpretations considered only the digital tools, not the physical and social environment of the learner, nor the pedagogical and technical skills used by learners to create and manage those tools. Again, the ‘environment’ metaphor was inadequate and misleading. The PLE was also, for the most part, a concept, not a technology, though efforts were made in some circles to create standards for mashing up those tools, most notably through work on ELF which was, by some, seen as the VLE of the future (Wilson, 2005, cited in Martindale & Dowdy, 2010), and a number of systems were built that were described as PLEs, but that were essentially another kind of institutionally managed server, much like Athabasca Landing, referred to previously.  A more promising set of standards that did focus on the development of standards-based widgets that could be assembled by individuals as well as within an LMS or other system (Wilson, Sharples & Griffiths, 2008), failed to gain enough momentum, despite endorsement of the widget specification by the W3 Consortium, and implementations within all major operating systems. Meanwhile, the term ‘PLE’ itself became such an amorphous concept that even conversations about it were difficult to sustain, let alone useful implementations.

In more recent years, the Educause organization has vigorously promoted the Next Generation Digital Learning Environment (NGDLE), which is essentially very similar in purpose and approach to the earlier ELF initiative, but that:

1)    Takes into account the possibility of learners assembling their own digital toolsets;

2)    Incorporates developments in analytics and artificial intelligence, and

3)    that is largely agnostic to standards used for its implementation, although it does recommend standards and protocols such as xAPI, LTI, learning record stores, and Caliper to help bind them together (Brown, Dehoney, & Millichap, 2015).

Combining the best ideas from service-based systems and work on PLEs, the initiative shows promise. While, once again, the ‘environment’ metaphor fails to extend into the actual spaces that it is intended to be deployed, the initiative is a genuine move beyond the teacher-centric, classroom-inspired models of the LMS and towards a student-oriented service-provision model. There are now some implementations of the concept. For example, the OERu aggregates a wide assortment of open source tools systems providing services such as discussion, microblogging, blogging, wikis, social bookmarking, and so on, that can be used independently by students or as part of the university’s own system (Lane & Good, 2019). While these are still largely perceived as an environment composed of environments, the potential for such a design approach is to free us from the traditional classroom metaphors of the LMS.

Institutional teaching beyond virtual environments

A distance learner’s environment is never digital, though digital tools and services can comprise important parts of it. A learning environment is not just comprised of physical or virtual structures but also the social, pedagogical, organizational, personal, and other dynamic elements that determine how the parts of the structure evolve and interact. It is not just physical matter, or virtual systems, but also the people and what they do together. It is not just how teachers teach, but how learners teach themselves, and teach one another, and are taught by the countless teachers who create the websites, interactions, tools, and structures of the broader internet, and the many teachers who inhabit their own physical spaces, from family members to people in the street. How, therefore, should teachers in institutions teach, when they are just parts of someone else’s environment, co-players in the process, and what kinds of digital tools and systems will be needed to support that?

Perhaps one of the reasons that it is too easy to fall into the trap of thinking of the digital tools and systems as an environment is that it they are  an obvious class of things around which to put a boundary. However, an infrastructure is not just the digital tools but also the human-enacted methods, rules, protocols, and standards that accompany it. It is not just what we use, but the ways that we use it.  It is natural to focus mainly on the software and hardware when designing an online system to support learning, and thus to come to think of it as providing the learning environment itself. If, instead, we remember that we are only building tools to use in the learner’s own environment, and that we are just providers or curators, not controllers or managers of that environment, then a critical and oft overlooked design principle becomes clear: that online students are the primary orchestrators of their learning rather than, as in the physical classroom, their teachers.

An integrated learning infrastructure should therefore not attempt to replicate the form and structure of a traditional classroom, nor should it solely support teachers in assembling the tools needed for their teaching. Instead, the focus – both digitally and pedagogically – should be on making it possible for learners to assemble the services into their environments themselves, in order to avail themselves of the support they need, when they need it, for the purposes they intend. The processes, methods, techniques, tools, and structures that students bring with them are at least as important as those created by their teachers. An integrated learning infrastructure needs to support these aspects at least as much as the interconnections between software tools. Again, it is necessary to think of the environment as considerably more than just a set of digital components that it uses but one that includes the people, the spaces they inhabit, and the things that they do.  Pedagogically as well as technically, there may be a need to support students in making the best use of all of that, for instance to search well, to find people that can help them to learn, to organize their own learning process, but such support Is, again, a service on which students may draw, not a teacher-determined requirement. And those pedagogies themselves need to adapt: for example, those that rely on rewards and punishments to enforce compliance must be excised, while those that provide learners with autonomy should be amplified. New pedagogies will be needed that acknowledge the many teachers in a learner’s environment, that help them to traverse the complexity of it, to leverage the advantages and to avoid the pitfalls. Teachers will need to let go, but stay close.

Tools that involve engagement with others – the means to share, the means to discuss, the means to work together, schedule meetings, and so on – are connection points in learners’ environments that cannot usually be completely controlled by any one of them, because of the need to at least agree protocols through which to engage and, in many cases, the systems which they will use to interact . One way to deal with this problem is to make a decision to use a  small range of tools, ideally in consultation with students. A better approach is to use tools that give students a choice of toolset, using protocols or standards such as SMTP, Jabber,  iCal, WebMention, ActivityStreams, or NNTP. However, few new standards have gained traction in recent years thanks to the dominance of closed social media monoliths intent on locking users in to their systems, so this may unnecessarily limit the range of systems that may be used. Another approach, commonly used in Connectivist approaches to learning, is to aggregate what learners provide themselves, using standards like RSS or Atom, or proprietary APIs offered by tool providers, or mailing systems to collect what students have shared elsewhere. If that is impossible, even simple copy-and-paste by human beings (students, teachers, or others) may be sufficient to connect multiple systems: not everything in an ILI needs to be implemented in software. For example, student blogs may be shared through flexible  technologies such as email and messaging apps, then copied by themselves or by their teachers into shared wikis. One  interesting benefit of such approaches is that they can support both diversity and manageability, inasmuch as the management burden may be shared by the participants rather than taken on by a single teacher or institution. Students may choose which tools they use, rather than having them chosen by the teacher. This is the principle used by Connectivist MOOCs (Downes, 2008), in which one site aggregates the shared artefacts created in many different learner-managed systems.

A learner’s environment consists of much more than the digital tools and systems offered by an institution. While, to a large extent, much of this environment may be unknowable to their designated teachers, there is much value to those who seek to support student learning in discovering how they are learning, and what constitutes their learning environment. Learning – the process, the tools, and the ways of learning, and not just the products – must be made visible if teachers, including other students, are to help learners to learn (Hattie, 2013). Much use can be made of pedagogical approaches such as shared learning diaries or blogging, and some careful use may even be made of automated systems that indicate presence, or that record traces of visits, as long as their role is to provide support for understanding student learning, and not to provide the teacher with means to control of the student. Beyond individual courses, there may be much pedagogical value in encouraging learners to share their learning experience through media such as blogs, microblogs, and other online tools, which may (as long as means are available for the student to control their privacy as needed) be aggregated and shared across their whole distributed, diffuse environment. Rather than replicating the necessarily closed and time-limited nature of the classroom, the artefacts of learning and the relationships that are developed in the process may persist indefinitely. Connectivist MOOCs provide a useful model for this. For example, Cormier (2014) talks of ‘Zombie MOOCs’ in which learning and interaction persist long after the course itself is over.

Bringing about such changes at an institutional level requires both bottom-up and top-down support. Teacher’s pedagogies are normally  more malleable than digital tools, because they can adapt rapidly to any tools: they are, in Brand’s terms (Brand, 1997), the ‘stuff’. However, they are therefore also the most constrained by the structures into which they must slot, and the least able to significantly impact things at structural level.  A single teacher, or even a small group of teachers pressing for change is therefore unlikely to sway either institutional policy or the design of the LMS because, as we have seen, one LMS must address the needs of all, so anything that changes it must suit everyone.

From the top down, replacing the LMS with an integrated learning infrastructure is a necessary step towards breaking out of the vicious loops that prevent the pedagogies and structures from evolving. At first, an ILI will naturally resemble the LMS it replaces, because its boundaries will continue to be largely determined by the less flexible layers above it: the institutional forms and structures such as courses, credentials, legislation, and teachers’ employment contracts. It is important to remember that the LMS was originally designed not just to replicate classroom behaviours but to fit into the larger, slower-changing structures and systems of institutions, and that significant changes in how we teach will not occur unless those structures and systems also evolve. They create the boundaries within which the ILI operates and, to a large extent, are not just containers of it, but part of it. An infrastructure is not just the digital tools but also the human-enacted methods, rules, protocols, and standards that accompany it. It is not just what we use, but the ways that we use it.  However, unlike the LMS, in an ILI those boundaries will be malleable. This opens up opportunities for the structure, skin, and services to in turn change.

The opportunities for change may not be taken, at first, at least in part because the signals (such as qualified students, their credentials, and so on) that pass in and out of the boundaries of the university will go to and from governments, employers, and other institutions that may not be prepared for radical change, even if the institution itself is committed to it. If, say, other institutions insist on grade point averages for standardized courses, then it will be difficult to completely avoid providing them, or something that is recognizably equivalent.  However, the adjacent possible empty niches (Kauffman, 2019) that an ILI supports will inevitably be filled by those who see the opportunities it entails, from courses whose lengths are pedagogically determined, to integration of lifelong and workplace learning, to new forms of credentials and learning. Perhaps, if enough institutions start to adopt such practices, we may break free of the insular single-institution model of education altogether. Out of this may grow a truly learner- and learning-driven future, in which learners draw on services from multiple educational providers, leading to a vast participative system in which institutions meld or blend to offer support for learning not just any time and any place, but every time and every place.

References

Blum, S. D., & Kohn, A. (2020). Ungrading: Why rating students undermines learning (and what to do instead). West Virginia University Press.

Bouygues, H. L. (2019). Does Educational Technology Help Students Learn? https://reboot-foundation.org/does-educational-technology-help-students-learn/

Brand, S. (1997). How buildings learn. Phoenix Illustrated.

Brown, M., Dehoney, J., & Millichap, N. (2015). The next generation digital learning environment. A Report on Research. ELI Paper. Louisville, CO: Educause April, 5(1), 1-13.

Cormier, D. (2014). Community learning – the zombie resurrection. http://davecormier.com/edblog/2014/05/25/community-learning-the-zombie-resurrection/

Downes, S. (2008). Places to Go: Connectivism & Connective Knowledge. Innovate, 5(1). http://www.innovateonline.info/pdf/vol5_issue1/Places_to_Go-__Connectivism_&_Connective_Knowledge.pdf

Dron, J.  (2006). Any color you like, as long as it’s Blackboard®.  In. Hawaii: AACE.

Dron, J., & Anderson, T. (2014). Teaching crowds: Learning & Social Media. AU Press. http://teachingcrowds.ca

Dron, J. (2016). P-learning’s unwelcome legacy. TD Tecnologie Didattiche, 24(1), 72-81. http://www.tdjournal.itd.cnr.it/article/view/891

Dron, J. (2021). Educational technology: what it is and how it works. AI & SOCIETY, 1-12. https://doi.org/10.1007/s00146-021-01195-z

Greene, E. B. (1928). The relative effectiveness of lecture and individual reading as methods of college teaching. Genetic Psychology Monographs.

Hattie, J. (2013). Visible Learning: A Synthesis of Over 800 Meta-Analyses Relating to Achievement. Taylor & Francis.

Kauffman, S. A. (2019). A World Beyond Physics: The Emergence and Evolution of Life. Oxford University Press.

Kelly, K. (2009). Triumph of the Default. http://www.kk.org/thetechnium/archives/2009/06/triumph_of_the.php

Keuning, H., Jeuring, J., & Heeren, B. (2018). A systematic literature review of automated feedback generation for programming exercises. ACM Transactions on Computing Education (TOCE), 19(1), 1-43.

Kohn, A. (2011). The case against grades. Educational Leadership, 69(3), 28-33.

Kohn, A. (2015). Four Reasons to Worry About “Personalized Learning”. http://www.alfiekohn.org/blogs/personalized/

Lane, D. C., & Good, C.  (2019). OERu’s delivery model for changing times: An Open Source NGDLE.  In World Conference on Online Learning, Dublin, Ireland. https://docs. oeru. org/s/fXQk2rJbzWCk8ia.

Martin, F., Chen, Y., Moore, R. L., & Westine, C. D. (2020). Systematic review of adaptive learning research designs, context, strategies, and technologies from 2009 to 2018. Educational Technology Research and Development, 68(4), 1903-1929.

Martindale, T., & Dowdy, M. (2010). Personal learning environments. Emerging technologies in distance education, 177, 193.

O’Neill, R.V., DeAngelis, D.L, Waide, J. B., & Allen, T. F. H. (1986). A Hierarchical Concept of Ecosystems. Princeton University Press.

Postman, N. (1998). Five things we need to know about technological change. Denver, Colorado, 28.  https://student.cs.uwaterloo.ca/~cs492/papers/neil-postman–five-things.html

Protopsaltis, S., & Baum, S. (2019). Does Online Education Live Up to its Promise? A Look at the Evidence and Implications for Federal Policy.

Ryan, R. M., & Deci, E. L. (2017). Self-determination theory: Basic psychological needs in motivation, development, and wellness. Guilford Publications.

Tichavsky, L. P., Hunt, A. N., Driscoll, A., & Jicha, K. (2015). “It’s Just Nice Having a Real Teacher”: Student Perceptions of Online versus Face-to-Face Instruction. International Journal for the Scholarship of Teaching and Learning, 9:2. http://digitalcommons.georgiasouthern.edu/ij-sotl/vol9/iss2/2/

Verbert, K., Duval, E., Klerkx, J., Govaerts, S., & Santos, J. L. (2013). Learning analytics dashboard applications. American Behavioral Scientist, 57(10), 1500-1509.

Wilson, S. (2008). Patterns of personal learning environments. Interactive learning environments, 16(1), 17-34.

Wilson, S., Sharples, P., & Griffiths, D.  (2008). Distributing education services to personal and institutional systems using Widgets.  In Proc. Mash-Up Personal Learning Environments-1st Workshop MUPPLE 8 (pp. 25-33).

Yen, C.-J., Tu, C.-H., Sujo-Montes, L. E., Harati, H., & Rodas, C. R. (2019). Using personal learning environment (PLE) management to support digital lifelong learning. International Journal of Online Pedagogy and Course Design (IJOPCD), 9(3), 13-31.

A few thoughts on learning management systems, and on integrated learning environments and their implementation

Why do we build digital learning systems to mimic classrooms?

It is understandable that, when we teach in person, we have to occupy and make different uses of the same or similar environments like classrooms, labs, workshops, lecture theatres, and offices. There are huge financial, physical, and organizational constraints on making the environment fit the task, so it would be madness to build a whole new classroom every time we wished to run a different class.

Online, we could build anything we like

But why do we do the same when we teach online? There are countless tools available and, if none are suitable, it is not too hard to build them or modify them to suit our needs. Once they are built, moving between them just takes a tap of a screen or the click of a mouse. Heck, you can even occupy several of them at once if you have a decent monitor or more than one device.

So why don’t we do this?

Here are a few of the more obvious reasons that using the perfect app for the context of study rarely happens:

  • Teachers’ lack of knowledge of the options (it takes time and effort to discover what’s available).
  • Teachers’ lack of skill in using them (most interesting tools have a learning curve, and that gets steeper in inverse proportion to the softness and diversity of the toolset, so most teachers don’t even know how to make the most of what they already have).
  • Lack of time and/or money for development (a real-life application is what it contains, not just the shell that contains it, and it is not always as easy to take existing stuff and put it in a new tool as it might be in a physical space).
  • Costs and difficulties in management (each tool adds costs in managing faults, configuration, accounting for use, performance, and security).
  • Cognitive load involved for learners in adapting to the metaphors, signposts, and methods needed to use the tool itself.

All of these are a direct consequence of the very diversity that would make us want to use different apps in the first place. This is a classic Faustian bargain in which the technology does what we want, and in the process creates new problems to solve.  Every virtual system invents at least some of the dynamics of how people and things interact with it and within it. In effect, every app has its own physics. That makes them harder to find out about, harder to learn, harder to develop, costlier to manage, and more difficult to navigate than the static, fixed facilities found in particular physical locations. They are all different, there are few if any universals, and any universal today may become a conditional tomorrow. Gravity doesn’t necessarily work the same way in virtual systems.

image of a pile of containersAnd so we get learning management systems

The learning management system (LMS) kind of deals with all of these problems: poorly, harmfully, boringly, and painfully, but it does deal with them. Currently, most of the teaching at Athabasca University is through the open source Moodle LMS, lightly modified by us because our needs are not quite like others (self-pacing and all that). But Moodle is not special: in terms of what it does and how it does it, it is not significantly different from any other mainstream LMS – Blackboard, Brightspace, Canvas, Sakai, whatever.

Almost every LMS essentially automates the functions, though not exactly the form, of traditional classrooms. In other parts of the world people prefer to use the term ‘managed learning environment’ (MLE) for such things, and it is the most dominant representative of a larger category of systems usually described as virtual learning environments (VLEs) that also includes things like MOOs (multi-user dungeons, object oriented), immersive learning environments, and simpler web-based teaching systems that replicate aspects of classrooms such as Google Classroom or Microsoft’s gnarly bundle of hastily repurposed rubbish for teaching that I’m not sure even has a name yet. Notice the spatial metaphors in many of these names.

Little boxes made of ticky tacky

The people who originally designed LMSs back in the 90s (I did so myself) based their designs on the functions and entities found in a traditional university because that was their context, and that was where they had to fit. Metaphorically, an LMS or MLE is a big university building with rather uniform classrooms, with perhaps a yard where you can camp out with a few other systems (plugins, LTI hooks, etc) that conform to its requirements and that are allowed in to classrooms when invited, and a few doors and gateways (mainly hyperlinks) linking it circuitously or in jury-rigged fashion to other similarly weakly connected buildings (e.g. places to register, places to seek support, places to talk to an advisor, places to complain, places to find books, and so on). It doesn’t have metaphorical corridors, halls, common rooms, canteens, yards, libraries or any of the other things that normally make up a physical university. You rarely get to even be aware of other classrooms beyond those you are in. Some people (me in a past life) might give classrooms cute names like ‘the learning cafe’ but it’s still just another classroom. You teleport from one classroom to the next because what happens in corridors (really a big lot of incredibly important pedagogically useful stuff, as it happens) is not perceived by the designers as a useful classroom function to be automated or perhaps, more charitably, they just couldn’t figure out how to automate that.

Reified roles

It’s a very controlled environment where everyone has a programmatically enforced role (mostly reflecting traditional educational roles), that may vary according to the room, but that are far less fluid than those in physical spaces. There are strong hierarchies, and limited opportunities for moving between them. Some of those hierarchies are new: the system administrator, for instance, has way more power than anyone in a physical university to determine how learning happens, like an architect with the power to move walls, change the decor, add extensions, and so on, at will. The programmers of the system are almost god-like in their command of its physics. But the ways that they give teachers (or learning designers, or administrators) control, as designers, directors, and regulators of the classroom, are perhaps the most pernicious. In a classroom a teacher may lead (and, by default, usually does). In an LMS, a teacher (or someone playing that role) must lead. The teacher sees things that students cannot, and controls things that the students may not. A teacher configures the space, and determines with some precision how it will be used. With a lot of effort and risk, it can be made to behave differently, but it almost never is.

Functions are everything

An LMS is typically built along functional lines, and those functions are mostly based on loose, superficial observations of what teachers and students seem to do in physical classrooms. The metaphorical classrooms are weird, because they are structured by teaching (seldom learning) function rather than along pedagogical lines: for instance, if you want to talk with someone, you normally need to go to a separate enclosed area inside the classroom or leave a note on the teacher’s desk. Same if you want to take a test, or share your work with others. Another function, another space. Some have many little rooms for different things. Lectures are either literally that (video recordings) or (more usefully, from a learning perspective), text and images to be read on screen, based on the assumption that the only function of lectures is information transmission (it is so very, very much not – that’s its least useful and least effective role). There’s seldom a chance to put even put up your hand to question something. Notices can usually only be pinned on the wall by teachers. Classroom timetables are embodied in software because of course you need a rigid and unforgiving timetable in a medium that sells itself on enabling learning anywhere, any time. Some, including Moodle, will allow you to break up the content differently, but it’s still another timetable; just a timetable without dates. It’s still the teacher who sets the order, pacing and content.

Robot overlords

It’s a high-tech classroom. There are often robots there that are programmed to make you behave in ways determined by those higher in the hierarchy (sometimes teachers, sometimes administrators, sometimes the programmers of the software). For instance, they might act as gatekeepers that prevent you from moving on to the next section before completing the current one, or they might prevent you submitting work before or after a specified date. They might mark your work. There are surveillance cameras everywhere, recording your every move, often only accessible to those with more powerful roles (though sometimes a robot or two might give you a filtered view of it).

Beginnings and ends

You can’t usually go back and visit when your course is over because someone decided it would be a good idea to set opening and closing enrolment dates and assumed that, when they were done, the learning was done (which of course it never is – it keeps on evolving long after explicit teaching and testing occurred). Again, it’s because physical classes are scheduled and terms come to an end because they must be, not because it makes pedagogical sense. And, like almost everything, you can override this default, but hardly anyone ever does, because it brings back those Faustian bargains, especially in manageability.

Dull caricatures of physical spaces

Basically, the LMS is an automated set of metaphorical classrooms that hardens many of the undesirable by-products of educational systems in software in brain-dead ways that have little to do with how best to teach, and that stretch the spatial metaphors that inform it beyond breaking point. Each bit of automation and each navigational decision hardens pedagogical choices. For all the cozy metaphors, programmers invent rather than replicate physics, in the process warping reality in ways that do no good and much harm. Classrooms solved problems of physics for in-person teaching and form part of a much larger structure that has evolved to teach reasonably well (including corridors, common rooms, canteens, and libraries, as it happens). Their more visible functions are only a part of that and, arguably, not the main part. There is much pedagogy embedded in the ways that physical universities, whether by accident or design, have evolved over centuries to support learning in every quadrangle and nook of a coffee shop. LMSs just focus on a limited subset of teaching roles, and empower the teacher in ways that caricature their already excessive dominance in the classroom (which only occurred because it had to, thanks to physics and the constraints it imposed).

LMSs are crap, but they contain recognizable semblances of their physical counterparts and just enough configurability and flexibility to more or less work as teaching tools, a bit, for everyone, almost no matter what their level of digital proficiency might be. They more or less solve the Faustian bargains listed earlier, but they do so by stifling what we wanted and should have been able to do in the first place with online tools, in the process creating new and quite horrific problems, as well as demolishing most of what makes physical universities work in the first place. It never has been true that virtual learning environments are learning environments – they are only ever parts of them – and there are places to escape from them, such as the Landing, other virtual systems, or even just plain old email, but then all those Faustian bargains come back to haunt us again. There has to be a better way.

Beyond the LMS

Cognisant of the issues, Athabasca University is now some way down the path to developing its own distinctive solutions to these problems, in a multi-year multi-million-dollar initiative known as (following the spatial metaphor) the Integrated Learning Environment (ILE). The ILE is not an application. It is an umbrella term for a lot of different, usually independent systems working together as one. Though some of the most interesting opportunities are still only loosely imagined, perhaps because they cause problems that are fiendishly hard to solve (e.g. how can we integrate systems that we build ourselves without creating risks for the rest of the ILE, and what happens when they need to be maintained?) a lot of progress is being made on the non-teaching foundations on which the rest depends (student admin systems, support tools, procedures, etc), as well as on the most visible and perhaps the biggest of its parts, BrightSpace, a proprietary commercial LMS that is meant to replace Moodle, for no obvious pedagogical or technical reasons (it’s no better). It might make economic sense. I don’t know, but I do know that open source software typically costs a fair bit to own, albeit because of the things that make it a much better idea (freedom, flexibility, ownership, etc). There is probably a fair bit of time and money being spent with Desire2Learn (makers of Brightspace) on the things that we spent a fair bit of time and money on many years ago to make Moodle a bit less classroom-like. The choice no doubt has something to do with how reliably and easily it can be made to work with some of the other proprietary commercial systems that someone has decided will make up the ILE. It bothers me greatly that we are not trying hard to choose open source solutions, for reasons that will become clearer in the rest of this post. However, (pedagogically speaking) all the mainstream LMSs are much of a muchness, making the same mistakes as one another in very similar ways, so it probably won’t wreck too much of what we already do within Moodle. But, on its own, it won’t move us much further forward and we could do it better. That’s what the ILE is supposed to do – to make the LMS just a part of a much larger teaching environment, intimately connected with the rest of what the university does for or with students, and extensible with new and better ways of learning, teaching, and assessing learning.

picture of lego bricksLego bricks make poor metaphors

When we were first imagining the ILE, though the approach was admirably participative, engaging much of the university community, I was very worried by the things we were encouraged to focus on. It was all about the functionality, the usability, the design, the tools, the pedagogies, the business systems that supported them. Those things matter, for sure, and should be not be ignored, but they should and will change and grow all the time: in fact, part of the point of building this thing is to do just that. Using the city metaphor, pretty much all that we (collectively) considered were the spaces (the rooms, mainly), and the stuff that goes on inside them, much like LMS designers thought of universities as just collections of classrooms in which teaching functions were performed. Space and stuff are, not uncoincidentally, exactly what Stewart Brand identified long ago as inevitably being the fastest-changing, most volatile parts of any town or city (after site, structure, skin, and services). I’ve written a fair bit on the universality of this principle across all systems. It’s a solid structural principle that applies as much to ecosystems and educational systems as to cities. As Brand observes himself, drawing from O’Neill et al (1986), the larger, slower-changing elements of any system affect the smaller, faster-changing more than vice versa. This is for much the same reasons that path dependencies set in. It’s about the prior providing the context for what follows. Flexible things have to fit into the gaps left by less flexible, older, pre-existing things. In physical spaces, of course these tend to be bigger and/or slower, but the same is true in virtual spaces, where size seldom matters that much, but hardness (inflexibility, brittleness) really does. Though lip service was paid to the word ‘integrated’ in our discussions,  I had the strong feeling that the kind of integration we had in mind was that of a Lego set. In fact, I think we were aiming to find a ‘Lego Athabasca University’ set, with assembly instructions and a picture on the box. The vendors who came to talk with us made much of how effectively they could do that, rather than how effectively they could make it possible for others to do that.

Metaphors matter. Lego bricks have to fit together tightly, in pre-specified ways, especially if you are following a plan. If you want to move them around, you have to dismantle a bit of the structure to fit them in. It’s difficult to integrate things that are not bricks, or that are made by different toy companies to work in different ways. At best you get what Brand calls ‘magazine architecture’, or ‘no road’ architecture, beautiful, fit for purpose, intricate and solid, but slow to learn. Lego is not a terrible way to build, compared with buying everything pre-assembled, but it could be improved.

Signals and boundaries

Drawing inspiration from John Holland’s brilliant last work, Signals & Boundaries, I tried to make the case that, instead, we should be focusing on the boundaries (the interfaces between the buildings and the rest of the city), and the signals that pass between them (the people, the messages, etc, the forms they take and how they move around). In Brand’s terms, I wanted us to be thinking about skin and services, and perhaps even structure, though site – Athabasca University – was a given. Though a few people nodded in agreement, I think it mainly fell on deaf ears. We wanted oven-ready solutions, not the infrastructure to enable those solutions. Though the city metaphor works well, because we are talking about human constructions, others would result in similar ways of thinking: cells in bodies, organisms in ecosystems, brains, termite mounds, and so on. All are organized by boundaries (at many levels of hierarchy) and the signals that pass between them.

The Lego set metaphor – whether deliberately or not – seems to have prevailed for now. A lot of old buildings are being slated for demolition and a lot of new virtual buildings are now being erected as part of this development, many of them chosen not because of problems with existing buildings but so that they can more easily connect together and live in the same cloud. This will very likely work, for now, but it is not cheap and it is not flexible, especially given the fact that most of it is not open so, like a rental property, we are not allowed to fix things, add utilities, change the walls, etc, and we are wholly dependent on the landlords being nice to us and each other (knowing that some – ahem, Microsoft – have a long history of abusing their tenants). Those buildings will age. We will find them cramped. Some will age faster than others, and will have to be modified to keep up, perhaps at high cost. Companies renting them might go out of business or change their terms so we might have to demolish the buildings and rent/make new ones. We will be annoyed at how they do things, usually without asking us. We will hate the landlords who dictate what we can do and how we can do it, and who will keep upping the rent while not doing what we ask. We will want more, and the only way to get it will be to build extensions, buy new brick sets, if it is not enough to pay someone to remodel the interiors (and it won’t be). Of course, because most of the big structural elements will not be open source, we will not be able to do that ourselves.

What the ILE really should be

The ILE is, I think, poorly named, because it should not be an environment at all. Following the building metaphor, the ILE is (or should be) more like the system that connects a lot of buildings, bringing them together into a coherent, safe, livable community. It’s infrastructure and services; it is the roads, the traffic signals, the doors, the sidewalks, the water pipes, the waste pipes, the electricity, the network cables; it is the services – fire, police, schools, traffic control, etc; it is all the many rules, standards, norms and regulations that make them work together to help make an environment in which people can live, work, play, and grow. It’s part of the environment – the part that makes it work – but it is not the environment itself. The environment itself is Athabasca University, not just the tools, processes, and systems that support its functions. That includes, most importantly, the people who are part of the university, or who are visitors to it, who are not just users of the environment or dwellers in its walls, but who are or should be the most significant and visible parts of it, just as trees are part of the environment of forests, not users of the forest. Those people live in physical as well as other virtual environments (social media, Word documents, websites, etc) that the ILE can connect together too, to make them a part of it, so the spatial metaphor gets weird at this point. The ILE makes environmental boundaries fuzzy, permeable, and shifting. It’s not an ILE, it’s an ILI – an integrated learning infrastructure.

If we focused on the connections and interfaces, and on how information and processes need to pass across them, and if we thought hard about the nature of those signals, then we could build a system that is resilient, that adapts, that lasts, that grows, that evolves, with parts that we can seamless replace or improve because the interfaces – the building facades, the mains pipes, the junction boxes, etc – will mostly stay the same, evolving slowly as they should. This is about strategy, not planning,  a way of thinking about systems rather than a sequence of things to do.

Some of the key people involved in the process realize this. They are talking about standards, protocols, and projects to build interfaces between systems, and imagining future needs, though they are inevitably distracted by the process of renting Lego bricks, so I am not sure how much they will be able to stay focused on that. I hope they prevail over those who think they are building a set of classrooms and tightly connected admin offices out of self-contained interlocking bricks because our future depends on getting it right. We are aiming to grow. It just takes one critical piece in the Lego building to fail to support that, and the rest falls apart like a… well, like a pile of bricks.

References

Brand, S. (1997). How buildings learn. Phoenix Illustrated. https://www.penguinrandomhouse.ca/books/320919/how-buildings-learn-by-stewart-brand/9780140139969

Holland, J. H. (2012). Signals and Boundaries: Building Blocks for Complex Adaptive Systems. MIT Press.  https://mitpress.mit.edu/books/signals-and-boundaries

O’Neill, R.V., DeAngelis, D.L, Waide, J. B., & Allen, T. F. H. (1986). A Hierarchical Concept of Ecosystems. Princeton University Press. http://www.gbv.de/dms/bs/toc/025157787.pdf

Postman, N. (1998). Five things we need to know about technological change. Denver, Colorado, 28.  https://student.cs.uwaterloo.ca/~cs492/papers/neil-postman–five-things.html

The NGDLE: We Are the Architects | EDUCAUSE

A nice overview of where the NGDLE concept was earlier this year. We really need to be thinking about this at AU because the LMS alone will not take us where we need to be. One of the nice things about this article is that it talks quite clearly about the current and future roles of existing LMSs, placing them quite neatly within the general ecosystem implied by the NGDLE.

The article calls me out on my prediction that the acronym would not catch on though, in my defence, I think it would have been way more popular with a better acronym! The diagram is particularly useful as a means to understand the general concept at, if not a glance, then at least pretty quickly…

ngdle overview

Address of the bookmark: https://er.educause.edu/articles/2017/7/the-ngdle-we-are-the-architects

Originally posted at: https://landing.athabascau.ca/bookmarks/view/2752680/the-ngdle-we-are-the-architects-educause

Learnium

Learnium is yet another attempt to overlay a cloud-based social medium on institutional learning, in the same family as systems like Edmodo, Wikispaces Classroom, Lore, GoingOn, etc, etc. I deliberately exclude from this list the far more excellent, theoretically grounded, and innovative Curatr, as well as dumb bandwagoners like – of all things – Blackboard (not deserving of a link but you could look up their atrocious social media management tools if you want to see how not to do this).

Learnium has a UK focus and it includes mobile apps as well as institutional integration tools. It looks slick, has a good range of tools, and seems to be gaining a little traction. This is trying to do something a little like what we tried to do with the Landing, but it should not be confused with the Landing in intent or design philosophy, notwithstanding some superficial similarities. Although the Landing is often used for teaching purposes, it deliberately avoids things like institutional roles, and deliberately blurs such distinctions when its users make use of them (eg. when they create course groups). It can be quite confusing for students expecting a guided space and top-down structure, and annoying if you are a teacher trying to control the learning space to behave that way, but that’s simply not how it is designed to work. The Landing is a learning space, where everyone is a teacher, not an institutional teaching space where the role is reserved for a few.

Learnium has a far more institutionally managed, teacher/course-oriented perspective. From what I can tell, it’s basically an LMS, cut down in some places, enhanced in its social aspects. It’s closer to Canvas than Moodle in that regard. It might have some value for teachers that like the social media tools but that dislike the lack of teacher-control, lack of privacy, deeply problematic ethics, and ugly intrusions of things like Facebook, and who do not want the cost or hassle of managing their own environments.  It is probably a more congenial environment for social pedagogies than most institutional LMSs, allowing learning to spread beyond class groups and supporting some kinds of social networking. There is a lot of scope and potential for vertical social networks like this that serve a particular kind of community in a tailored fashion. This is very much not Facebook, and that’s a very good thing.

But Learnium is an answer to the question ‘how can I use social media in my courses?’ rather than ‘how can social media help to change how people learn?’ It is also an answer to the question of ‘how can Learnium make money?’ rather than ‘how can Learnium help its users?’ And, like any cloud-based service of this nature (sadly including Curatr), it is not a safe place to entrust your learning community: things like changes to terms of service, changes to tools, bankcruptcy ,and takeovers are an ever-present threat. With the exception of open systems that allow you to move everything, lock stock and barrel, to somewhere else with no significant loss of data or functionality, an institution (and its students) can never own a cloud-based system like this. It might be a small difference from an end user perspective, at least until it blows up, but it’s all the difference in the world.

Address of the bookmark: https://www.learnium.com/about/institutions/

Original page

The LMS as a paywall

I was writing about openness in education in a chapter I am struggling with today, and had just read Tony Bates’s comments on iQualify, an awful cloud rental service offering a monolithic locked-in throwback that just makes me exclaim, in horror, ‘Oh good grief! Seriously?’ And it got me thinking.

Learning management systems, as implemented in academia, are basically paywalls. You don’t get in unless you pay your fees. So why not pick up on what publishers infamously already do and allow people to pay per use? In a self-paced model like that used at Athabasca it makes perfect sense and most of the infrastructure – role-based time-based access etc – and of course the content already exists. Not every student needs 6 months of access or the trimmings of a whole course but, especially for those taking a challenge route (just the assessment), it would often be useful to have access to a course for a little while in order to get a sense of what the expectations might be, the scope of the content, and the norms and standards employed. On occasion, it might even be a good idea to interact with others. Perhaps we could sell daily, weekly or monthly passes. Or we could maybe do it at a finer level of granularity too/instead: a different pass for different topics, or different components like forums, quizzes or assignment marking. Together, following from the publishers’ lead, such passes might cost 10 or 20 times the total cost of simply subscribing to a whole course if every option were purchased, but students could strategically pick the parts they actually need, so reducing their own overall costs.

This idea is, of course, stupid. This is not because it doesn’t make economic and practical sense: it totally does, notwithstanding the management, technical and administrative complexity it entails. It is stupid because it flips education on its head. It makes chunks of learning into profit centres rather than the stuff of life. It makes education into a product rather than celebrating its role as an agent of personal and societal growth. It reduces the rich, intricately interwoven fabric of the educational experience to a set of instrumentally-driven isolated events and activities. It draws attention to accreditation as the be-all and end-all of the process. It is aggressively antisocial, purpose-built to reduce the chances of forming a vibrant learning community. This is beginning to sound eerily familiar. Is that not exactly what, in too high a percentage of our courses, we are doing already?

If we and other universities are to survive and thrive, the solution is not to treat courses and accreditation as products or services. The ongoing value of a university is to catalyze the production and preservation of knowledge: that is what we are here for, that is what makes us worthwhile having. Courses are just tools that support that process, though they are far from the only ones, while accreditation is not even that: it’s just a byproduct, effluent from the educational process that happens to have some practical societal value (albeit at enormous cost to learning). In physical universities there are vast numbers of alternatives that support the richer purpose of creating and sustaining knowledge: cafes, quads, hallways, common rooms, societies, clubs, open lectures, libraries, smoking areas, student accommodation, sports centres, theatres, workshops, studios, research labs and so on. Everywhere you go you are confronted with learning opportunities and people to learn with and from, and the taught courses are just part of the mix, often only a small part. At least, that is true in a slightly idealized world – sadly, the vast majority of physical universities are as stupidly focused on the tools as we are, so those benefits are an afterthought rather than the main thing to celebrate, and are often the first things to suffer when cuts come along. Online, such beyond-the-course opportunities are few and far between: the Landing is (of course) built with exactly that concern in mind, but there’s precious little sign of it anywhere else at AU, one of the most advanced online universities in the world.  The nearest thing most students get to it is the odd Facebook group or Twitter interaction, which seems an awful waste to me, though a fascinating phenomenon that blurs the lines between the institution and the broader community.

It is already possible to take a high quality course for free in almost any subject that interests you and, more damagingly, any time now there will soon be sources of accreditation that are as prestigious as those awarded by universities but orders of magnitude cheaper, not to mention compellingly cut-price  options from universities that can leverage their size and economies of scale (and, perhaps, cheap labour) to out-price the rest of us. Competing on these grounds makes no sense for a publicly funded institution the role of which is not to be an accreditation mill but to preserve, critique, observe, transform and support society as a whole. We need to celebrate and cultivate the iceberg, not just its visible tip. Our true value is not in our courses but in our people (staff and students) and the learning community that they create.

The EDUCAUSE NGDLE and an API of One's Own (Michael Feldstein)

Michael Feldstein responds on NGDLEs with a brilliant in-depth piece on the complex issues involved in building standards for online learning tool interoperability and more. I wish I’d read this before posting my own most recent response because it addresses several of the same issues with similar conclusions, but in greater depth and with more eloquence, as well as bringing up some other important points such as the very complex differences in needs between different contexts of application. My post does add things that Michael’s overlooks and the perspective is a little different (so do read it anyway!), but the overlapping parts are far better and more thoroughly expressed by Michael.

This is an idea that has been in the air and ripe for exploitation for a very long time but, as Michael says in his post and as I also claim in mine, there are some very big barriers when it comes down to implementing such a thing and a bunch of wicked problems that are very hard to resolve to everyone’s satisfaction. We have been here before, several times: let’s hope the team behind NGDLE finds ways to avoid the mistakes we made in the past.

Address of the bookmark: http://mfeldstein.com/the-educause-ngdle-and-an-api-of-ones-own/

Niggles about NGDLEs – lessons from ELF

Malcom Brown has responded to Tony Bates and me in an Educause guest post in which he defends the concept of the NGDLE and expands a bit on the purposes behind it. This does help to clarify the intent although, as I mentioned in my earlier post, I am quite firmly in favour of the idea, so I am already converted on the main points. I don’t mind the Lego metaphor if it works, but I do think we should concentrate more on the connections than the pieces. I also see that it is fairly agnostic to pedagogy, at least in principle. And I totally agree that we desperately need to build more flexible, assemblable systems along these lines if we are to enable effective teaching, management of the learning process and, much much more importantly, if we are to support effective learning. Something like the proposed environment (more of an ecosystem, I’d say) is crucial if we want to move on.

But…

It has been done before, over ten years ago in the form of ELF, in much more depth and detail and with large government and standards bodies supporting it, and it is important to learn the lessons of what was ultimately a failed initiative. Well – maybe not failed, but certainly severely stalled. Parts persist and have become absorbed, but the real value of it was as a model for building tools for learning, and that model is still not as widespread as it should be. The fact that the Educause initiative describes itself as ‘next generation’ is perhaps the most damning evidence of its failure.

Elves

Why ELF ‘failed’

I was not a part of nor close to the ELF project but, as an outsider, I suspect that it suffered from four major and interconnected problems:

  1. It was very technically driven and framed in the language of ICTs, not educators or learners. Requirements from educators were gathered in many ways, with workshops, working groups and a highly distributed team of experts in the UK, Australia, the US, Canada, the Netherlands and New Zealand (it was a very large project). Some of the central players had a very deep understanding of the pedagogical and organizational needs of not just learners but organizations that support them, and several were pioneers in personal learning environments (PLEs) that went way beyond the institution. But the focus was always on building the technical infrastructure – indeed, it had to be, in order to operationalize it. For those outside the field, who had not reflected deeply on the reasons this was necessary, it likely just seemed like a bunch of techies playing with computers. It was hard to get the message across.
  2. It was far too over-ambitious, perhaps bolstered by the large amounts of funding and support from several governments and large professional bodies. The e-learning framework was just one of several strands like e-science, e-libraries and so on, that went to make up the e-framework. After a while, it simply became the e-framework and, though conceptually wonderful, in practical terms it was attempting far too much in one fell swoop. It became so broad, complex and fuzzy that it collapsed under its own weight. It was not helped by commercial interests that were keen to keep things as proprietary and closed as they could get away with. Big players were not really on board with the idea of letting thousands of small players enter their locked-in markets, which was one of the avowed intents behind it. So, when government funding fizzled out, there was no one to take up such a huge banner. A few small flags might have been way more successful.
  3. It was too centralized (oddly, given its aggressively decentralized intent and the care taken to attempt to avoid that). With the best of intent, developers built over-engineered standards relying on web service architectures that the rest of the world was abandoning because they were too clunky, insufficiently agile and much too troublesome to implement. I am reminded, when reading many of the documents that were produced at the time, of the ISO OSI network standards of the late 80s that took decades to reach maturity through ornate webs of committees and working groups, were beautifully and carefully engineered, and that were thoroughly and completely trounced by the lighter, looser, more evolved, more distributed TCP/IP standards that are now pretty much ubiquitous. For large complex systems, evolution beats carefully designed engineering every single time.
  4. The fact that it was created by educators whose framing was entirely within the existing system meant that most of the pieces that claimed to relate to e-learning (as opposed to generic services) had nothing to do with learning at all, but were representative of institutional roles and structures: marking, grading, tracking, course management, resource management, course validation, curriculum, reporting and so on. None of this has anything to do with learning and, as I have argued on many occasions elsewhere, may often be antagonistic to learning. While there were also components that were actually about learning, they tended to be framed in the context of existing educational systems (writing lessons, creating formal portfolios, sequencing of course content, etc). Though very much built to support things like PLEs as well as institutional environments, the focus was the institution far more than the learner.

As far as I can tell, any implementation of the proposed NGDLE is going to run into exactly the same problems. Though the components described are contemporary and the odd bit of vocabulary has evolved a bit, all of them can be found in the original ELF model and the approach to achieving it seems pretty much the same. Moreover, though the proposed architecture is flexible enough to support pretty much anything – as was ELF – there is a tacit assumption that this is about education as we know it, updated to support the processes and methods that have been developed since (and often in response to) the heinous mistakes we made when we designed the LMSs that dominate education today. This is not surprising – if you ask a bunch of experts for ideas you will get their expertise, but you will not get much in the way of invention or new ideas. The methodology is therefore almost guaranteed to miss the next big thing. Those ideas may come up but they will be smoothed out in an averaging process and dissenting models will not become part of the creed. This is what I mean when I criticize it as a view from the inside.

Much better than the LMS

If implemented, a NGDLE will undoubtedly be better than any LMS, with which there are manifold problems. In the first place, LMSs are uniformly patterned on mediaeval educational systems, with all their ecclesiastic origins, power structures and rituals intact. This is crazy, and actually reinforces a lot of things we should not be doing in the first place, like courses, intimately bound assessment and accreditation, and laughably absurd attempts to exert teacher control, without the slightest consideration of the fact that pedagogies determined by the physics of spaces in which we lock doors and keep learners controlled for an hour or two at a time make no sense whatsoever in online learning. In the second place, centralized systems have to maintain an uneasy and seldom great balance between catering to every need and remaining usably simple. This inevitable leads to compromises, from small things (e.g. minor formatting annoyances in discussion forums) to the large (e.g. embedded roles or units of granularity that make everything a course). While customization options can soften this a little, centralized systems are structurally flawed by their very nature. I have discussed such things in some depth elsewhere, including both my published books. Suffice to say, the LMS shapes us in its own image, and its own image is authoritarian, teacher-controlled and archaic. So, a system that componentizes things so that we can disaggregate any or all of it, provide local control (for teachers and other learners as well as institutions and administrators) and allow creative assemblies is devoutly to be wished for. Such a system architecture can support everything from the traditional authoritarian model to the loosest of personal learning environments, and much in between.

Conclusion

NGDLE is a misnomer. We have already seen that generation come and go. But, as a broad blueprint for where we should be going and what we should be doing now, both ELF and NGDLE provide patterns that we should be using and thinking about whenever we implement online learning tools and content and, for that, I welcome it. I am particularly appreciative that NGDLE provides reinvigorated support for approaches that I have been pushing for over a decade but that ICT departments and even faculty resist implacably. It’s great to be able to point to the product of so many experts and say ‘look, I am not a crank: this is a mainstream idea’. We need a sea-change in how we think of learning technologies and such initiatives are an important part of creating the culture and ethos that lets this happen. For that I totally applaud this initiative.

In practical terms, I don’t think much of this will come from the top-down, apart from in the development of lightweight, non-prescriptive standards and the norming of the concepts behind it. Of current standards, I think TinCan is hopeful, though I am a bit concerned that it is becoming over-ornate in its emerging development. LTI is a good idea, sufficiently mature, and light enough to be usable but, again, in its new iteration it is aiming higher than might be wise. Caliper is OK but also showing signs of excessive ambition. Open Badges are great but I gather that is becoming less lightweight in its latest incarnation. We need more of such things, not more elaborate versions of them. Unfortunately, the nature of technology is that it always evolves towards increasing complexity. It would be much better if we stuck with small, working pieces and assembled those together rather than constantly embellishing good working tools. Unix provides a good model for that, with tools that have worked more or less identically for decades but that constantly gain new value in recombination.

Footnote: what became of ELF?

It is quite hard to find information about ELF today. It seems (as an outsider) that the project just ground to a halt rather than being deliberately killed. There were lots of exemplar projects, lots of hooks and plenty of small systems built that applied the idea and the standards, many of which are still in use today, but it never achieved traction. If you want to find out more, here is a small reading list:

http://www.elframework.org/ – the main site (the link to the later e-framework site leads to a broken page)

http://www.elframework.org/projects.html  – some of the relevant projects ELF incorporated.

https://web.archive.org/web/20061112235250/http://www.jisc.ac.uk/uploaded_documents/Altilab04-ELF.pd – good, brief overview from 2004 of what it involved and how it fitted together

 https://web.archive.org/web/20110522062036/http://www.jisc.ac.uk/uploaded_documents/AltilabServiceOrientedFrameworks.pdf  – spooky: this is about ‘Next Generation E-Learning Environments’ rather than digital ones. But, though framed in more technical language, the ideas are the same as NGDLE.

http://www.webarchive.org.uk/wayback/archive/20110621221935/http://www.elearning.ac.uk/features/nontechguide2 – a slightly less technical variant (links to part 1, which explains web services for non-technical people)

See also https://web.archive.org/web/20090330220421/http://www.elframework.org/general/requirements/scenarios/Scenario%20Apparatus%20UK%205%20(manchester%20lipsig).doc and https://web.archive.org/web/20090330220553/http://www.elframework.org/general/requirements/use_cases/EcSIGusecases.zip, a set of scenarios and use cases that are eerily similar to those proposed for NGDLE.

If anyone has any information about what became of ELF, or documents that describe its demise, or details of any ongoing work, I’d be delighted to learn more!