Sunday, July 15, 2007

The Quest for Sustainability in Open Courseware


Note: This article was originally posted in the Connect section on the Educause Web site, at:
http://connect.educause.edu/blog/pault/thequestforsustainability/44767. 
However, this address has since become inaccessible, so the post has been reproduced here as an archive with the same date and approximately the same time.


I've been reflecting recently on the subject of open courseware and, more specifically, OpenCourseWare following the keynote for the Sakai conference in Amsterdam delivered confidently and enthusiastically by Hal Abelson (a podcast is available). In this post I'll briefly recap some of the core aspects as I understand them and then go on to explore this area, based on personal experiences and ideas I've been formulating at Oxford.

Abelson took a broad view, inviting the audience to go back 25 years and defined programming as a "novel formal medium for expressing ideas." Against that, he got us to consider the aspirations and expectations that we might have had then, encapsulating this in 3 predictions for 25 years thence (i.e. today):
  • a global encyclopaedia
  • TCP/IP global
  • collaborative educational resources
It's the third that has yet to be properly delivered. Starting from consideration of why not, he then developed the rationale leading to the MIT OpenCourseWare (OCW) initiative and the more recent Creative Commons Learn (ccLearn).

Abelson described OCW and ccLearn as means to building infrastructure for sharing academic pursuits, covering platforms and materials in Sakai, policy structure and media structure, designed in such a way as to protect academic values. The need to beware certain kinds of commercial activities was drilled into the audience: such concerns are, he argued, keen on monopolising and overcharging us. So, in the face of impending monopoly, it was argued that we need OCW, shared repositories etc, in order to be taken seriously at national and international levels. The IPR issue highlights a tension between the commercial and academic world. He urged everyone that we shouldn't leave it to the publishers to control, and by way of illustration mentioned that universities can have a policy on publication that insists on the right to retain rights and publishers should be sought that allow reasonable IPR.

Enter Creative Commons' ccLearn:


Our goal is to make material more "interoperable," to speed up the virtuous cycle of use, experimentation and reuse, to spread the word about the value of open educational content, and to change the culture of repositories to one focused on "helping build a usable network of content worldwide" rather than "helping build the stuff on our site.

It's new to me and one month on I've subsequently tried to find out more. I certainly haven't searched far, but ccLearn still seems largely hidden, with little information available: someone who hears about it might well type cclearn in Google and would find cclearn.com, the 'Center for Creative Learning,' which has also taken the domain cclearn.org. I found it difficult to come across much of substance regarding ccLearn - just a few snippets, e.g. a mention on Stanford's Center for Internet and Society. Of course, as it's a Creative Commons project, you could go to the creativecommons.org site, but when I entered cclearn site:creativecommons.org in Google only one match was returned! At least it informs us that they now have an Executive Director - Ahrash N. Bissell - congratulations to him :-)

The presentation itself flowed swiftly with ease until ... there was a big anticlimax at the end when the economic realities became evident - in Q&A at the end he admitted that the average cost of preparing an MIT course the OCW way is around $15,000-$20,000, mainly down to legal concerns apparently.

Indeed in a subsequent session, 'Open courseware, pedagogy , Social Practices and Tools,' which elaborated on OCW initiatives, major problems with the current OCW were identified:
  • too expensive to create OCW sites
  • little or no automation
  • no connection to CLE
  • only large institutional commitment can get OCW off the ground
OCW is a meritorious activity and undoubtedly makes a major contribution to making more visible the academic enterprise - the Webometrics 'World University Rankings' provide some indication of this with MIT sitting on top of the table (whereas Oxford lies many places beneath). It can be argued that these are very limited measures, but Web visibility really does count.

Given that it's worthwhile, but costly, how might there be economic sustainability? One might look for inspiration to open source software (OSS) generally and follow the example of seeking revenue from support, certification etc., but I expect this has already been covered. More specific to the educational context, the Open courseware session expressed the hope that the next generation of OCW, dubbed OCW2, will reduce cost by employing graduate students, trained to understand licensing, and enabling them to share in the academic sphere. To enable this, they are looking at incentive structures, trying to get early buy in. The graduate helpers are called Digital Scribes whose engagement can work positively to foster "co-creation" and "communities," but I think graduates may well swap and change how they earn enough to get by, so can't always be depended on. We also heard that from another point of view, OCW may be regarded as filling out the long tail of publishing (a phrase coined by Chris Anderson), as illustrated by Amazon, which is able to sell at least one copy of every book, no matter how obscure, thus offering a chance to support specialisms (J.R. Hartley would be pleased!) and I guess Lulu is another good illustration. However, overall, I'm not convinced this will be much better.

So what would this small person from a small island suggest as an alternative approach?

Allow me to start with a quote from one of last year's extraordinary debates on the governance of Oxford University. It comes from Donald Fraser, Professor of Earth Sciences, who as reported in proceedings from Congregation , 14 November, 2006 stated:
Dynamic knowledge-based businesses are moving away from large, centrally administered monoliths, towards small, self-organising entrepreneurial cells, flexibly connected and practically self determining—just look at the campus models of companies like 3M, Google and Apple.

What does that mean to me as someone who works in academic support? The message I read (and readily agree with) is that academics rather than administrators are the ones who, along with their colleagues and peers, are in the best position to determine what they should do with their academic activities - in terms of how it can help them, their department, their field of study and their students. In the context of the debate as a whole, he was arguing against the motion because it contained proposals that were seen as increasing central control over the academics in ways that would threaten their independence and autonomy. From this, I infer that essentially that academic endeavour starts internally and is facilitated by an inter-networking mode of operation. If you look at the origin and flow of ideas, it often starts wthin one individual, spreads to a group and then more widely. It's a fact not just of research, but of teaching and of any other activity. Institutions need to support this as best they can, particularly as individuals are becoming increasingly mobile, moving from one institution to another.

This view of academic freedom doesn't deny the institution and its overall mission, but it does ask for a light touch, in terms of how academic enterprise is directed and also in terms of general bureacracy, particularly the legal aspects. I guess this is one of the major issues of OCW and I wonder if OCW2 really lessens this. I think a basic lesson to take from the governance debate (I'm not sure I could grapple with many of the subtleties) is that we should seek first to clarify principles: the professor is the academic authority who should drive the decision-making subject to the authorisation of the institution. In order for this to work effectively, the authorisation should be devolved, which is actually the traditional way in which Oxford works. If it's not suitably devolved, then you get a lot of overhead, so that institutional approval becomes necessary for very small steps, making things very expensive.

Such a devolved view can then transfer much of the responsibility to individuals, requring them to focus especially on basically two main issues:
  • appropriate use of content that you haven't produced yourself
  • deciding on the rights you wish to grant to content you have produced
If these issues are addressed as early as possible in the course creation lifecycle - by determining what's needed in the way of permissions and what should be granted in the way of rights - then that should save a lot of resources later on. With the right training, by the time materials are published the first time in a course management system, the main licensing issues and policy should already be resolved so that when it comes to making available as open courseware, the main effort is technical. This is dependent, I think, on authorisation at the highest level established as early as possible, ideally at the outset, so that it is quickly devolved. The kinds of authorisation I have in mind is a policy document on the kinds of licensing that are permitted, how the University is identified with each publication, specifically giving academic members the rights to publish according to Creative Commons licenses subject to various terms and conditions. Gaining authorisation itself may not be easy, though, as the institution will likely require strong arguments as to the benefits of making content free to use and repurpose - ICT staff may already have had a taste of this in trying to persuade their institutions to let them release software under an open source license.

Assuming processes can be put in place, what does this mean for implementation?

The OCW presentations I've attended have conveyed the sense that OCW is a long way from just open educational content - I certainly got that impression from the Educause '06 presentation Open Sharing, Global Benefits - The OCW consortium where open educational resources - were defined in terms of digitised materials offered freely and openly to use and reuse for teaching, learning and research; whereas open courseware are specific kind of educational resource materials, which have to be organised around a course, though the duration is open. There's a lot of emphasis on process and, in particular, OCW requires that content must be IP-cleared: every contribution gets passed through and checked - sometimes it is removed or replaced where it is felt that copyright has not been granted on at least some content. When I stepped back to reflect on openness in open courseware, I could see quite a few severe hurdles to surmount, some of which seemed unnecessary. Such a heavyweight approach has led to some consideration of sustainability in terms of a few institutions managing the processes, hosting OCW content, and selling this as a service: Wolfgang Greller sees this is an opportunity for OpenLearn, the OU's version of OCW.

However, I have reservations about the hosting provision at such institutional level through third parties and, in any case, my view is that we are dealing essentially with another output, one that results from existing internal processes to which most resources have already been devoted. Rather, institutional ownership can be expressed naturally through their own LMS, which can provide many organisational benefits, not least a single point of access to all study resources for students and for external examiners. However, If we are to support academics individually as originators of content, then the LMS system needs to support personalisation, a flexible environment in which to organise and publish. Indeed, I feel that the way Oxford is run in a devolved and self-organising way points to more organic and sustainable means that make sense particularly with the host of Web2.0 technologies are available. Hence, I now feel more confident that an LMS can provide valid linkage between personalisation and open courseware, as intimated in my poster at last year's conference in Dallas.

I think we should try to envision how it would work for an academic. I imagine a Professor accessing a LMS and going straight to their personal area, in which they have inter alia options to create, review and share content. For Oxford users it means using MyWebLearn, which makes available all the tools necessary to author a course. Sharing the material can be carried out literally in a few steps:
  1. Log in.
  2. Go to the resource you wish to make public
  3. Click on the link 'View Access' at the bottom of the page.
  4. In the following page go to the pull-down menu 'Allow..' and select 'Public' to 'look at' this page.
  5. Click on the [Add] button to enact.
This simple mechanism has already been used to some extent in WebLearn, evident in Google with a few thousand resources (pages) indexed compared with fewer than a hundred pages from another institutional VLE with the same name! However, this process only enables the materials to be put in the open. From the academic's perspective, there needs to be added to this the means for specifying the licensing. Assuming a suitable policy and process were in place, then options could easily be added. Overall it needs to be very easy to use, ideally as easy as contributing to a blog.

On the other side of the coin, materials published this way as courses need to satisfy certain organisational and structural requirements - the content should be sourced from departmental areas, which need to be planned and designed into the system. Also, to be discoverable they need to be indexed with suitable metadata; and interfaces need to be provided that pull together all the relevant information in a meaningful way. We can achieve this by mapping to institutional structures, e.g. the LMS can automatically insert meta data about department, so that subsequently presenting the courses on offer as a whole, can be achieved by aggregation, say. Here I think we can learn from Warwick blogs, an institutional blog hosting service in which staff and students are able to write freely and connect with others. However, they have linked in with their institutional NDS LDAP directory, so that you can browse blogs based on department and even module of study. WebLearn already uses the institutional map in that it is hierarchical in structure, with the top two levels controlled centrally as far as departments and colleges. However, once at that level, areas are managed locally, i.e. content creation has been decrentralised, allowing natural growth.

The issue of quality control should already be handled in the processes of preparing the courses at the institution; what is being provided is largely a snapshot of the materials that were used in live courses. Whatever the processes, I think it is important that the decisions about releasing such content are devolved as much as possible and that the mechanisms for effecting it are as easy as the illustration above. I understand that for OCW(2) processes are being developed for Sakai to make publication a smoother process, so perhaps the production of Creative Commons licensed content may be an option in future, though I wonder how devolved it is and whether it revolves around MyWorkspace. Also, until Sakai has hierarchy, in comparison the technicalities of generating such materials appear far easier in Bodington (and I suspect developing pipeline processes to go with them might be easier also).

If another editorial layer is needed, then that can emerge from peer networks. A number of years ago I came across the Hippias search engine, a service (now merged with Noesis) that as I recall had an editorial board of experts in Philosophy whose members each maintained their own Web sites. These sites contained links to other sites and the Hippias search engine would index all the pages at the end of these links, thereby building a trusted indexed collection. I think it's a very apposite illustration of how you can combine devolved human quality control with automation.

This is obviously work in progress and much is still open to debate, but from the view I've described above, I think the focus should very much be with the academics, devolving much of the decision-making and supporting them as appropriate. Technically, this means Web2.0-like approaches should be incorporated and so I expect many elements of ccLearn could play a major role in facilitating institution-oriented OCW.
I hope to talk more about personalisation and Web 2.0 in future posts...

Monday, July 02, 2007

All aboard? Reflections on the 7th Sakai conference, Amsterdam


Note: This article was originally posted in the Connect section on the Educause Web site, at:
http://connect.educause.edu/blog/pault/allaboardreflectionsonthe/44626. 
However, this address has since become inaccessible, so the post has been reproduced here as an archive with the same date and approximately the same time.


Oxford made a decision in Autumn 2006 to migrate to the Sakai VLE with the announcement of the Tetra collaboration. Since the completion of the academic year, we've been able to focus more on the task in hand. For myself, I decided the best way to quickly gain a feel for Sakai was to attend a Sakai gathering and conveniently the 7th Sakai Conference was recently held in Amsterdam, the first time the conference had been held outside the United States. I was primarily interested in sessions that addressed system migration, deployment and support, but also keen to hear about pedagogy and usability, leaving it to my colleagues to cover the more technical development aspects. I wanted to know what approaches were adopted to move to Sakai: organisation, resources, timescales, etc.

So was it a case of all aboard...?

Silver Shadow in port at Amsterdam

Above is the luxury cruise liner, Silver Shadow, which was waiting for passengers to board. It was right next to the Moevenpick Hotel, the conference venue. In fact, it's designed to accommodate a little under 400 guests, about the number of participants at the conference, which rather suggests a dream future venue ... :-)

It's taken two or three weeks for my impressions to settle - I found the three days of the conference quite intense and took copious notes. I can say straightaway, however, that I felt there was generally a good sense of community, with a very constructive outlook across the various constituent communities, ranging from development through to pedagogy and research. Sessions were usually informative and presented well; there was a real sense of purpose and commitment It was consistent with what I've observed on the Tetra developers mailing list: although I'm not been involved in any Java coding myself, I have been seen how the Sakai developers have provided very helpful responses to the various queries raised by Bodington developers seeking to incorporate key functionality in Sakai. Furthermore, when some of these ideas were presented by my colleagues, Adam and Matthew, in their presentation on importing Bodington tool, they were greeted very positively - there is a willingness to learn. So I broadly concur with the encouraging sentiments expressed by Michael Feldsteain in his 'State of the Union' blog post.

There remain many questionmarks as expressed Ian Reid, whose responses were not so rosy: in wrapping it up, he perceived a number of weaknesses and for him fundamental questions remained unanswered. I can at least answer his first point about the product: there are certainly large scale deployments - e.g. at Indiana and Michigan. Further, many of the other points, such as the technical bias, are well known and as far as I can tell they are being actively addressed.

I have quite a number of concerns myself and among my colleagues may be the one who is most reluctant to migrate from our present WebLearn, based on Bodington, perhaps largely because I have spent so much time with it and naturally can get attached. My first query is what kind of system is Sakai? Is it largely an open source replacement for Blackboard or WebCT? At the culmination of the procurement process at Oxford in 2001/2002, we were left with a head-to-head between Blackboard and Bodington. There was a free vote and Bodington won very easily, largely because Bodington offered flexibility: in the use of terminology, in how it allowed areas to be set up, in who could do what in these areas, and in how users could navigate freely around the system. One could use it to augment existing teaching or research arrangements with little effort. WebLearn has subsequently grown organically - from the handful of resources in December 2002 to its present state of about 60,000 resources manually created and managed by thousands of users (staff and students) in the various colleges and departments. At the same time, Bodington also has many weaknesses - it's rather long in the tooth and has often been described as "clunky" - many of the tools are looking very dated and making changes can be very laborious.

Sakai was felt to be the most promising way forward, but as it stands there are serious limitations in its design. The name of Michigan's deployment itself hints at one of these 'CTools,' rather indicating a technical focus: indeed much of the talk at the conference was 'tools' oriented, but during the past year or two, in WebLearn, we've deliberately tried to move away from 'how does tool X work' to 'how to carry out activity Y [using the tools available]' with a recent project looking at activity-based use cases for WebLearn. Also, the course subscription model in Sakai will not be not sufficient (in theory, Oxford's undergraduates are at liberty to attend any lecture at the University); the role-based access control is more coarse-grained (in fact, Bodington doesn't have any fixed roles - they can be defined via group memberships), and the overall organisation of materials lacks hierachy. There are many other smaller issues - e.g. what about those horrible Sakai URLs?

There are many concerns, but there is reassuringly intense activity to address them and this is leading to mutual enrichment. So, a lot of discussion has flowed on the topic of groups; a new hierarchy service for Sakai might have a name component that will enable nice URLs etc. I also saw some good examples of how requirements are driving the development; how development goes through a proper processes of evaluation and many other encouraging signs, such as the use of the term of Collaborative Learning Environment (CLE), getting away from the systems-oriented language.

So, largely reassured, at this stage my biggest concern is more in terms of timescales and resources regarding a full deployment of Sakai: it's a question of when than if. Looking around, it seems fitting then to note Stanford's announcement on 21 June:

After a year-long pilot, Sakai went into full production at Stanford today, fully replacing our legacy home-grown system. We've taken a long, careful path toward deployment to assure a seamless transition to the new system. It is localized, integrated and well tested, and today we flipped the switch. This is a big achievement for us, fulfilling the commitment we made to ourselves, and to our collaborators at Indiana, Michigan and MIT three years ago when we started this project.

So it can be done, but a long road lies ahead and if we are to achieve this at Oxford for everyone's benefit, we really shall need all aboard!