We want to develop a framework for thinking about OER (开放教育资源), along a few different dimensions… The most important is the purpose (目的), and the use (用法?) - of course, they don't need to be the same. So the first question is, why do people produce OER, or why do governments, Hewlett foundation, universities, etc, fund the production of OER? What's their goal? And how do they actually get used? And if we know the answer to this, what are the implications for what kind of material we need, the license, format, quantity, quality, production process etc? (Implicit in this statement is the understanding that by failing to specify what purpose a given project wants to achieve, it is impossible to properly analyze whether it has been successful or not, and it is hard to tell how the project should be designed. In terms of TLCP, I get the feeling that the government wants to achieve “everything”, which is impossible, and makes for a muddled approach.) 

    We can separate purposes into 4 (let me know if you have any other ideas of improving this framework): process, direct use, transparency and reuse. 

    Process (I discussed this name with my friend for quite a while, because he said it didn't work in Chinese. We came up with “Production” instead)

    This covers all the benefits accrued from the process of producing the OER in itself, before the OER has even been finished or published. This seems especially relevant to TLCP, and is something that I hadn't even included before! The goal of TLCP is improving the quality of undergraduate teaching. The source is existing materials, but rather than snapshot, MoE wants the material to change: set up teaching teams, get more teachers (senior profs) involved in teaching undergrad, reflecto on pedagogics, content, update everything. Thus, the process has a positive effect on the teacher/teachers receiving the funding directly - which would still exist even if they never shared the product of their work with others.

    Even in the West, the process of production can have both positive and negative effects. (In this case, we are mostly talking about times when teachers adapt existing material for distribution). If integrated production is the case, how does it change the way a teacher teaches… some US profs have reported that they feel less “free”, less open to experiment etc. In a similar way, when I give a presentation and I know that it will be recorded and made available publicly, there are certain things I will not say, which I might have said if it was just the 30 people listening to my lecture…


    In this case, we are looking for requirements for designing the process itself, rather than the resulting materials. There doesn't seem to have been much research on how the OER funding/creation process should be designed so as to maximize positive impact and minimize negative impact. Are you aware of studies from China where they have looked at the effect of a teacher's own teaching from getting a TLCP grant. (Ie. Not whether he was great from the beginning, but whether getting a grant made him better)… There is a case study from a university in Lanzhou that talks about this (TLCP在教学工作中的作用)

    These courses don't need an open license or even to be distributed… A research question is how the fact that the courses will be distributed, changes the process and the outcome for the teachers involved?  

    Direct use

    Target group: could be students, could be teachers or profs - but who want to learn themselves, not use it to teach others. Could also be life-long learners, ie anyone in society who would like to learn something. For students, it could be related to their major, but it could also be completely unrelated.

    Goal: To be able to learn something in an independent fashion. Without having the guidance of a teacher etc. Ie, I want to learn about economics to run my business. I go to a course, I begin reading materials, I answer some multiple choice tests that immediately tell me whether it's right or wrong, I play with some simulations, I improve my knowledge etc.

    Requirements: Very different from snapshot -  we need material that is specifically designed for the learning. Its also hard to design such a self-contained course without knowing anything about the target group, so usually some assumptions are made about the target group, although the course is open to anyone. A good deal of pedagogics (教学方法) and curriculum design is needed, can use theories from distance ed, even though it is different, because here there is no teacher or human interaction (usually). It might help if there are interactive parts (simulations, quizzes etc), but these are difficult to develop, and can also make it less easy to remix, translate etc the material (although not necessary).

    The material should also be complete, or at least all the essential material that it refers to should also be available for free online. (This is different from MIT courses, which often refer to articles and books that are not freely available). Since most university courses do not fulfill this requirement, it means that the course will have to be adapted quite strongly for the web. If the course if from a distance university, it might need a lot less adaptation (like OU UK). Many of these courses are designed from scratch, however.

    Quality is very important, and it is assessed with an entire course as a unit: Is this course good or not?

    Classic example: Carnegie Mellon Open Learning Initiative. (There are _many_ articles abt this, I can find them). Their purpose is to replace the gigantic first year introductory courses in US universities - like economics 101, phyisics 101… Use a lot of learning sciences and pedagogy to design, strong interactive element, Has research that proves that it is very efficient. Might only work for certain kinds of courses. Also very tightly integrated, free to use, but cannot install on own server, easily translate etc. Courses with this purpose do not necessarily need an open license to fulfill their primary purpose (although an open license is always nice, and can create unforeseen opportunities for reuse etc).

    When it comes to research, I bet there is a lot of research on how to create these courses in an effective manner - for example for use in corporate training (paid by company, for workers). The only change when it becomes open is that it is free… needs high investment, and hard to make sustainable. 


    The idea behind this is to “look through other people's windows and see what they are doing”. Ie, you yourself do not necessarily want to learn directly from the course, nor do you want to directly reuse it (copying and editing parts of it, etc), instead you just want to get some inspiration, curious, etc.

    Important inspiration for my thinking here comes from two blogposts that Mike Caulfield wrote.

    Target groups; Two main components. The first is other teachers who can get inspiration… This seems to be a major component of TLCP. Students can also use it of course - to get ideas about how courses are built up, what books to read. In the MIT evaluation report they have many examples, for example about incoming students using it to select major, or even teachers using it to better coordinate between each others courses.

    An additional component that I have not seen mentioned much (if at all) In the literature, is the idea of using it for comparative research. I think this has lot's of  potential, but no organization is likely to fund work with this specific purpose, so it is a “use case” that has to “piggyback” on the first task mentioned.

    Requirements: This requires material that reflects as closely as possible what actually happens in the classroom, what is handed out to the students etc, The snapshot model. Quality is not so important, in fact the more material the better! The material doesn't need to be complete (and will not be, since some material has to be excluded for copyright and privacy reasons, and unlike for other “goals”, this does not require replacement material to be created etc). An open license is only necessary for possible translation (OCW is a good example - there is very little reuse, but it has been translated to several language, and there is some evidence of teachers for example in China using MIT to compare their own courses etc).

    Do you know anything about the comparative case studies between MIT OCW and Chinese university courses? According to CORE, 250 such book-length reports have been published! This is a great example.

    Economics: Arguably the impact of such courses is low (not sure if we can write that - cannot necessarily find support in the literature, my personal opinion), so it might be hard to secure large funding. This means that the ideal is to make the publication of such material as low-overhead as possible. The material already needs to be prepared and taught, the challenge is to put it online with as little as possible extra cost. An ideal thus is to integrate the production of new course material into the production of OER (for example simultaneous publishing to Blackboard and EduCommons). I can find some articles about this.

    Classic cases: All OCWs and TLCP seem to fall under here. In addition, the  textbooks from India and Indonesia are also included.  


    Reuse or remix is kind of the “holy grail” of OER… but it's difficult to achieve.

    The basic concept is that you can take a whole or a part of an OER, edit it, mix it with something else, integrate it into something bigger, translate it, etc.

    Target group: Could be anything… The great thing is that you don't need to accurately predict who will need your material, because others can repurpose your material to fit their own needs. Could be someone taking material from an MIT course and inserting it into an open textbook that they are building, could be creating a new completely open course out of bits and pieces from several “half-complete” courses (like MIT and TLCP).

    Requirements: In this case, it's absolutely obligatory to use an open license, which allows remixing (ie not the Non-Derivative clause of Creative Commons). Since you don't know where the material might end up, a license with less restrictions is better - Share-Alike makes it harder to remix with other material, and Non-Commercial means that I cannot print a textbook and sell it in Africa. It's also important (and this is often ignored) to use file formats that can be easily edited. This means that PDFs are bad. RealMedia video is also very hard to edit. For pictures that are composite (composed of more than one layer, for example a picture with text on top) they should provide the original Photoshop etc file, so that others can easily change that text, without influencing the picture. For text, HTML, Word or OpenOffice, or even a PageMaker/inDesign file is good. (It's always better to make things available in a file format that can be edited with OpenSource software). For video that has been edited from many clips (ie not just a one hour shot of a talk, but like a documentary), the ideal would be that all the original video clips be made available, with the iMOvie etc file - so that other people can make different cuts. The same with composite sound files. (This can be compared to the idea in open source, that you not only give me the right, but also give me the “source file”… The source file of a picture is the not the picture itself, but what you used to make it.

    Quality is less important than for direct use,  and mainly focuses on whether it is easy to reuse them (see above license + format). Since it's hard to predict in what kind of situations the material will be used, objectively assessing quality is also very hard.

    Learning objects: this has been a long discussion about ways of creating objects that are encapsulated, and can easily be inserted into a learning management system like BlackBoard. Each object should also have lot's of metadata, so people can easily find it.  I think it's basically failed, and David Wiley has written a lot about this. It's hard to predict what kind of units people want, typing in the metadata is too much work, etc. Rather, just make everything available as HTML etc, so that it is full-text indexed by Google, and make it easy for people to take the units they want and edit them.

    “Light-weight remixing”: A way of “light-weight” remixing is constructing a curriculum by linking to individual parts of other courses. This has a number of advantages and disadvantages (this blog entry makes a good discussion: http://flexknowlogy.learningfield.org/2009/03/23/early-decisions-on-reuse-of-oer-copy-or-link/#more-613)… Basically, linking means there doesn't need to be a free license. However, each individual module needs to be individually adressable. Also the source needs to be stable so links don't break. You are unable to edit, translate, select pieces smaller than each page, and integrate deeply with other material.

    Examples: Unfortunately, so far there are few examples of this. There are examples of curricula that are constructed through light-weight remixing, for example David Wiley's course. I remember there was a presentation at Dalian last year about the difficulties in remixing. There might be more information out there. Also a book published by David Wiley and friends about OERs was partly composed of “remixed” parts from other places… Part of the problem is that most universities are not willing to do this - they are happy to produce new material, so that they can put their name on it, but they are less willing to take another universities' materials and improve on them…Personally I would love to see this with TLCP, this is why I keep saying the quality of the entire courses doesn't matter that much. Just see if you can find enough good quality materials from different courses, and put it together. Of course, until there is an open license, we can only do that through linking.

    An example of something I would like to do with remix is the books by Indira Gandhi Open University. They put a ton of books online that are very good, but they are scans of PDFs, and they are relevant to the Indian context. If I could, I would choose one or a few, try to extract the text with OCR, use volunteers to correct it, put it on a wiki, and use peer-production to improve it, and make it relevant to more cases. Right now, my hands are tied.

    There are great other examples of reuse online, like videos, music remixes etc.



    I also wrote a little bit about how OERs are produced in relationship to existing material. Not sure if it fits in anywhere. We could also discuss who produces them etc, but the most important factor is WHY. What's the purpose… Not to make an exhaustive list of all categories.

    source - OER that is based on an existing course, or teaching materials, and OER that is produced specifically for an open purpose.

    - Based on existing materials

          - Snapshot (快照), take what already exists and put it online. Might not be perfect, complete (because of Copyright problems) or adapted to online/distance learning, but it more or less accurately reflects what we are doing right now.

          - Modify, improve, adapt for an open context

          - A third, which is a bit of a mix of the two first ones: integrated production. This means that production of OER is integrated into the production of the normal courses… For example a teacher who is preparing a course for BlackBoard, can simultaneously select that some of the material will be made available to everyone, where there is no Copyright problem.

    - Create from scratch

          - More freedom to design so that it fits specific needs. But we must know exactly what those needs are. Also much harder to fund/make sustainable




    So, we have shown four possible “purposes” of designing OERs, in terms of how they will be used, and the different “requirements” these carry with them in terms of license, format, content, etc. Note that it is possible for a single project to cover several use cases, however if it tried to cover all, it will probably not cover any of them well… In addition, there are many unexpected use cases, and the use of as open a license, and as reusable file formats as possible, will enable use that was not anticipated. We have also provided classic examples of each of the use cases.

    We hope these distinctions will help clarify thinking about how to design new OERs, and also in deciding how to evaluate existing OER.



    This is the outline I made up before I began writing the text above.




    To the people who create the course

    Creating teaching team

    Reflecting on course content

    Incl best teaching methods and update course content

    To other people

    Inspire other people in the department (re case study Lanzhou)

    Direct use

    By self-learners (incl other students)

    Needs self-contained materials (like OLI), interactive etc


    Inspire other teachers

    Comparative researchers

    Ex: Indonesian/Indian textbooks



    Take parts of the material and move/edit

- License

    Must be open for reuse


- Type of material

    Transparency: snapshot, quantity

    Direct use: quality, interactivity, self-contained

    Reuse: license, file format

    来源: 快照, modified, 结合, from scratch