The new Pressbooks directory includes H5P search

The new Pressbooks directory has officially launched. I had a chance to see demos of this earlier this summer and was very impressed with the work that Pressbooks has done in creating a federated search across the Pressbooks EDU eco-system.

One of the features I want to make sure gets noticed is that, not only can you search for open textbooks and books in the directory, but also H5P activities. This is a massive boost to the H5P ecosystem as it now provides another central point to begin to find H5P activities that can potentially be reused (the others I know of are the eCampus Ontario H5P Studio site and the soon to be released (93% done) H5P OER Hub).

However, unlike the other repos of H5P content, the nice thing about the Pressbooks Directory H5P search is that it allows you to see the context that the H5P activity was used in. As we know from Wiley’s Reusability Paradox, the effectiveness of a reusable object relies on the context it is used in (along with an open license) and this new search helps provide some of that context by locating H5P content within an existing resource, giveng educators the ability to reuse not only the H5P interactive but also reuse the context (aka the book content) that wraps around that H5P activity. You get to see how others intended to use the resource and what learning it was meant to support.

For example, this H5P interactive video activity from An Interactive Introduction to Organismal and Molecular Biology on its own doesn’t mean all that much.

However, when you see it within the context of the entire page it comes from, you get a better and clearer sense as to the purpose of the H5P activity and can reuse as much of that surrounding chapter for context to make that H5P activity a much more effective learning object in your own context. Super useful.

Coming back to Pressbooks (the company), I cannot stress how valuable a contribution to the open education ecosystem this directory is, and Pressbooks deserves huge kudos for the work that they have done. At a time when we see plenty of examples of bad faith actors in the edtech space and can point to examples of corporate openwashing, Pressbooks has time and time again demonstrated that you can build a sustainable company by being both a partner and a contributor to the open education ecosystem. There are not many companies that I have worked with over the years that have consistently demonstrated such a deep commitment to the core values of open education as Pressbooks and other companies would do well to learn by their example on how to become an invaluable member of the open education community.


Pressbooks cloning now with added H5P goodness

Steel Wagstaff at Pressbooks has been talking to me for the past few months of a relatively new feature of Pressbooks where, when you clone a Pressbooks book it will now bring over all the H5P activities within that book.

While this feature has been available for Pressbooks network users, for us at BCcampus where we host our own instances of Pressbooks that is a bit behind the update schedule for Pressbooks network clients, we have had to wait for the feature as we worked our way through updating our Pressbooks network. Thanks to the work of the BCcampus Dev/ops team and Josie Gray, the update to our networks happened a few weeks ago and I was able to finally test out the H5P cloning feature on our instance.

Caveat: if you are self-hosting your own Pressbooks network, you need to have the H5P plugin installed and both the H5P plugin and Pressbooks plugin need to be up-to-date. At the time I write this that means H5P WordPress plugin ver 1.15.0 and PB version 5.13.0, although, as you will see below, during my testing I uncovered a bug in the H5P WordPress plugin that is scheduled to be fixed in the next release.

I tested the cloning feature using this great Body Physics textbook from Open Oregon that I knew had a lot of H5P interactions in it. Cloning the book is not that difficult (Pressbooks has a good step by step guide) and, with the exception of the issue below, worked quite well.

The cloning process itself does take some time, about 10 minutes to complete for this particular book. For testing I am focusing specifically on how the H5P content came over (as opposed to the PB content) and, for the most part, the cloning routine did a very good job of copying all the H5P content included in the book.  When I went into the H5P Content section, I see all the H5P interactions there indicating that everything came over in the clone.

Screenshot of H5P Content area in Pressbooks showing all cloned H5P activities

If I check out the H5P interactions in the book, I can see that the H5P interactions kept their original context within the book. That is, they appear in the same place in the cloned PB book as the original source book. Additionally, when I looked at the embed url for each H5P content type in the cloned book, I see that the H5P content now lives within my book separate from the cloned book. It has not simply been embedded in my book from the source book. A full copy has been created within my book so that I can modify and customize without affecting the upstream version. Screenshot of H5P element showing the URL of where the H5P element is located, confirming that the H5P element has fully copied over.In the future, it might be useful to see the copy as a fork of the original with links back to the original version somehow instead of a separate and discrete copy of the element. But that would require a lot of future development work. Put that on the wishlist for the future (and, imo, low priority as I think that, while forking is an elegant way to be able to trace derivatives to/from, it would be something that very few users would ever actually use. I’d like to be proven wrong with that assumption someday, but right now I don’t think there is the level of engagement with revising OER’s that warrant the development effort).

However, all is not perfect. I notice that the author attribution in the copied H5P elements is incorrect. I am listed as the author of all of them.

Screenshot show me attributed as author of all H5P interactions

One of the features I love about H5P is that you can add CC licenses to each individual H5P element, which makes the rules around sharing and attributing content much easier and, imo, is one of the big reasons why I think H5P is an OER platform as there has been care and consideration on the part of the H5P developers to add this as default functionality in H5P. But it is problematic if all the H5P content imported loses attribution information, and falsely attributes all the imported content to the person doing the importing.

At first I thought that it might be that the attributions were not set correctly in the source material, and, sure enough when I went back to the original book I did not see any visible license information where I would expect to see it.

Screenshot of H5P element showing missing license informtionSo maybe there is no actual attribution information included with the original content and that is why the author info in the copy reverted to me, the default author of the book?

I tested with another book with the same result – I was credited as the author with all the imported H5P elements. I contacted Steel at Pressbooks to double check that this was unexpected behaviour and we did a quick test together, creating a new book with a new H5P element with proper attribution and imported that into a new Pressbooks book. Again, the H5P element was attributed to the person importing the book and not the original author.

Steel confirmed that this was unusual behaviour and created an issue in the PB GitHub repo reporting the issue. After some BA work on the Pressbooks side it was discovered that the issue was upstream in the H5P WordPress plugin. A report was filed there and the developers of the H5P WordPress plugin have developed a fix that will be released in the next release of the plugin. So, with luck, this cloning attribution issue should be fixed soon.

Other than that issue, however, the cloning routine itself is very slick and sets the foundations to make the reuse and adaptation of H5P content in Pressbooks much easier.


When your project goes to the dogs

After something like 15 years of connecting both virtually and face-to-face, I am very happy to finally be working with Alan Levine (@cogdog) on a project. I needed some technical help installing and configuring three open source math based homework systems that are being evaluated as part of the BCcampus Open Homework System project. Even though these are not WordPress projects, I know there is more to the dog than WordPress, and approached him to see if he would be able to help out with the project. And I am very glad that he said yes as he has brought exactly what I needed to make this part of the project go smoothly.

His first order of business was to configure some web space for the project. Now, I would have expected that he would have sent an email with three IP url’s pointing to different installs (something like or some other user unfriendly url that we often have to work with when working with sandbox servers). Instead, Alan set up a one page landing page and worked with our internal network tech to configure subdomains for each of the installations so that the faculty testers would have some sane links to follow while they were testing the three systems.

Screenshot of the this is not a test test site for the BCcampus open homework systems

Not only does this provide a fine landing place for testers to have everything in one neat and tidy package, but Alan has gone one step further to turn this landing page into a project status page so that, at a quick glance, everyone involved with the project can see the status of the technical work he is doing in a very open and transparent way.

Screenshot showing example of project status

He has also set up a way for testers to report issues back to him for follow up using a simple Google Form. Anyone who has an issue during the testing period can fill the form out and send him a screenshot of the problem, collecting all the technical issues for each platform in a handy single spreadsheet that will be invaluable at the end when we are evaluating the testing results.

Along the way, you see little glimpses of the human touch that marks it as an Alan project. Like the disclaimer he included at the bottom of the landing page with an animated homage to Mission Impossible.

Screenshot of Mission Impossible self-destructing tape from 60's TV showI really appreciate little touches like this that add a bit of levity and fun to a project.

I also appreciate that Alan has not hesitated to jump in and make contact with the developers of the three applications we are testing. So far he has discovered at least 2 minor issues in the code for one of the applications and has been working with the developers to fix the code. That is value added. Regardless of whether or not we decided to use this particular piece of software, Alan has made the project better by uncovering these errors and working with the developers to fix them.

Working in open source has it’s own challenges. One of them being that you need to add value to the community to have a voice in the community. Build some goodwill with the development community and establish a relationship. This kind of direct communication with the developers to help them improve their code goes a long way in establishing the foundation for long term relationships which puts you in good stead with the community, and Alan does it very well.

I am really happy this project has gone to this particular dog.


Some strategies for the Open Homework Systems project

I’ve spent a good part of my summer work on the Open Homework System (OHS) project in research and information gathering mode, trying to set a solid groundwork for the next 2 years of the project.

In addition to working on a definition of what an OHS is, I’ve been tapping some people within the system on the shoulder to pull together a project advisory group, and have been meeting with many people, both within the BC post-sec system and in the wider open education community, about the project, gathering information to help inform some strategies for the OHS project. Thank you to all have graciously given time to talk with me this past summer.

From all of these conversations, report reading, and pulling from some experiences and lessons learned from past open projects I have worked on, I’ve come up with a list of strategies that I think can help guide our work on this project. This list will form much of the discussion at the first meeting I have with our project advisory group.

A commitment to Open Source and OER

As with all BCcampus projects, all code, content, applications, etc. created with this project will be released with open licenses to ensure wide adoptability and adaptability by all institutions within the BC post-secondary system and beyond.

One platform will not be enough

Early environmental scans are finding that there are nuanced and specific homework platforms catering to specific disciplines and/or signature pedagogies. Finding one system that can be reasonably expected to cover all disciplines is unrealistic.  We can reasonably expect that there will be homework systems that are geared to;

  • STEM areas (Science, Technology, Engineering and Math)
    • In particular, Math is an area that makes heavy use of homework systems as it is a discipline that, as a core pedagogical principle, is built upon repetitive practice and instant formative feedback.
    • Math can further be broken into sub-sections including Calculus, Geometry, Trigonometry and Algebra, each one of which may have distinct requirements.
  • Business
    • Some areas, such as Economics and/or Finance, are similar to Math in that they have built a pedagogical model around repetitive practice and formative feedback.
  • Arts & Humanities
  • Social Sciences
  • Health & Human Services
  • Trades

Multiple platforms are too many

We don’t have the resources to expand into multiple platforms for each discipline and sub-discipline area, therefore we will need to concentrate our efforts or risk diluting limited resources.

Focus on 2 areas: STEM, and one other

While we cannot cater to every different discipline, there are likely platforms that may broadly cover multiple disciplines within STEM areas. As for a second area, my own thinking is leading me down a path to consider platforms that lean towards supporting Social Sciences and/or Humanities as, outside of STEM, those are areas where we tend to see large enrollment first and second year courses (see next point). Additionally, I am seeing open education gaining a strong and growing base of support in areas like Digital Humanities, so there could be opportunities to leverage work happening in DH. Whatever the second area, I suggest we focus our efforts on 2 platforms: One that leans more towards the STEM areas, and one that caters to another discipline area, recognizing there will likely be overlap and/or gaps even with those very broad categories. This is an area where some further research into where expensive homework systems are in use could help narrow down the scope.

Focus on high enrollment courses/subject areas

Like the Open Textbook project, we should look at systems that support high enrollment introductory (1st and 2nd year) courses where these systems are in high use to maximize impact and student savings.

We should not create a new platform from scratch

There are many existing, viable options out there that we can likely contribute to and make stronger with our resources.

We should focus on content creation of collaborative quizzes, questions sets, activities, etc.

While technical development of platforms will likely be important (especially in the areas of user interface and sensible workflows for faculty and students), what will get faculty adopting and using these systems is content. Pre-built, vetted instructional content that they can plug into their course and go. So, we should put resources into the creation, curating and vetting of content that populates whatever OHS we decide to support.

We should be looking toward open digital courseware as the final destination

Open homework systems being a step along the gradient to get us to open digital courseware. Our pathway should be to start with open content (in the form of open textbooks), incorporate more self-assessment questions and interactive activities that provide instant feedback (open homework system), then onto more adaptive learning functionality (TBD, but needs to be considered as part of our OHS project), and develop/contribute to open source software projects that can move us closer to this vision of open digital courseware. This pathway aligns with the digital courseware definition from the Courseware in Context quality framework, and is a pathway that many commercial digital publishers are on. In the short term this means focusing our activities in the OHS project to align with existing open textbooks within our library to both leverage previous work (see next point) and to future proof those resources to make the next transition to full open digital courseware.

We should look to leverage existing open work within the system

In addition to open source software, our efforts around homework systems and, in particular, content development for this systems, should be tightly aligned with existing open content that has already been created through the open textbook project. The practice questions and examples we develop should tightly align with the learning outcomes of open textbooks that have already been created, or exist within the OTB collection, in order to increase their adoptability, create stronger open resources, and move them further along the path to open digital courseware.

Data collection and data analytics: FIPPA and beyond

In addition to fully FIPPA complaint systems, we should look at ways in which a homework platform and the data generated by a homework platform could be presented to learners and instructors in meaningful ways, which can then be used to inform learning pathways for specific learners. Many commercial systems include analytics engines and, in order to establish feature parity with commercial systems, we should include ways in which to ethically collect and use meaningful user interaction data and present to both learners & instructors for pedagogical purposes.


H5P as homework system

I left this as a comment over on JR Dingwell’s blog as a response to his post Textbooks, The Printing Press, The Internet, and OER Revisited but thought I would include it here as it reflects some of my thinking around one potential path I’d like to explore as part of the Open Homework System(s) project. You can read the entire post, but here’s a brief summary of the key points of his post that prompted my comment.

In his post, JR makes a number of very good points, including his desire to see the work happening in the open education space around homework systems as way to take better advantage of the pedagogical affordances of the internet, and provide something more than randomly generated “drill n’ kill” question sets. 

No pressure 🙂

But JR does provide some examples of work he has done on feedback loop type activities, including work he has done using H5P, which reflects some of my recent thinking around the OHS project. Here is my comment.

While I do see the value in having systems that randomly generated question & answer sets for disciplines that utilize practice heavily (like math and some other STEM areas where open source tools like WeBWorK are heavily utilized already), it’s tools like H5P and the close alignment of H5P interactions with content within an open textbook that get me quite excited.

Specifically, I’d like to see an extension to some of the Pressbooks-enabled open textbooks we (BCcampus) have created as part of the open textbook project, and am looking in our collection for some potential candidates to enhance with (human) developed H5P interactivity pieces. As you point out, I think that having these kind of closely aligned learning activities designed and developed specifically to support the content in the open textbook are useful ways in which we can both improve learning using the open textbook, and extend the adoptability of those textbooks. And, if we collaboratively create questions and activities with instructors using some sprint methods, a great way to build some digital skills capacity within our system among instructors while making stronger learning resources.

While I agree that this kind of development is not necessarily something that pushes the boundaries of open pedagogy and leverages new pedagogical potentials of the internet, it does slowly build out networked infrastructures and capabilities across the system. And I do see these types of H5P learning interactions as discrete shareable open learning objects (for lack of a better term). H5P does allow the sharing and embedding of learning interactions which is an affordance that the internet does make easier.

So, while not perhaps pushing the boundaries of internet-enabled pedagogies, I do see H5P as an important next step in the evolution of open textbooks to something closer to what our original vision was of open textbooks back in 2012 when I started working on the OTB project at BCcampus – a truly interactive learning experience for students.