EPUB for Education and Digital Textbooks

[AdamSobieski]

The EDUPUB alliance formed in 2013. Its work included discussion of: (1) an education profile, (2) education structural semantics, (3) scriptable components, (4) packaging and integration, (5) open annotation, (6) distributable objects, and (7) QTI/LTI/Caliper best practices.

Digital textbooks are an important digital publishing scenario. New specifications and standards can facilitate new features and new technologies for each and every digital textbook.

The EPUB3 Working Group is presently formulating its charter. I would like to ask about whether the EPUB3 Working Group might find these EPUB for education and digital textbook scenarios to be of some interest? I would also like to ask about how we might best advance the state of the art with respect to digital textbooks at the W3C?

[wareid]

Thanks for raising this! In writing the draft charter we looked at EPUB and all of the associated documents. Right now the charter focuses on bringing the EPUB 3.2 revision to rec track with room for further revisions.

This does not mean we don't have ideas for the associated specs, including EDUPUB. The main issue for us is that many of the satellite specs don't have the same implementation support as the core spec. We've decided for now to leave the satellite specifications for the community group, which as mentioned in the update from the steering committee will be the Publishing Community Group (which you can join here: https://www.w3.org/community/publishingcg/). The goal of the CG is to incubate ideas and develop them to a point where the recommendation track makes sense, so we think that it will be a good place for some of the satellite specs to be worked on.

[AdamSobieski]

Thank you. I joined the Publishing Community Group.

With respect to implementation support, in my opinion it might make a sense to, at a future point, distinguish between an EPUB Reading System and an EPUB Digital Textbook Reading System. That is, not every EPUB Reading System need also be an EPUB Digital Textbook Reading System.

Towards incubating ideas and developing them to a point where the recommendation track makes sense, what would those here involved in Publishing@W3C think about organizing a new Task Force at the Publishing Community Group on these digital textbook topics with that Task Force having a deliverable report which: (1) discusses and potentially updates the aforementioned EDUPUB ideas, and (2) presents new ideas for digital textbooks and educational technology?

[iherman]

@mteixeira-wwn @rakutenjeff I wonder how to record this issue for the CG so that it isn't lost? This repository is concentrates on the WG charter, i.e., it may have to be closed, but I do not want the original comment of @AdamSobieski to become lost...

[iherman]

As discussed on the Steering Committee, moving this issue to the Publishing CG issue list

@mteixeira-wwn

[llemeurfr]

There is a new interest around us (i.e. EDRLab) for structural semantics applicable to textbooks, the target audience being print disabled people.

We feel that EDUPUB tried to kill too many birds with one stone. Interactive features should be kept out of publications (including them in publications made the whole thing too hard to implement), they should be implemented in specific apps.

The start of the discussion was that digital textbooks are usually created out of a mix of InDesign pages and diverse other sources. Such "print first" workflows are totally missing the point re. structural semantics. But for print disabled students we need highly structured content (-> highly structured exercises). Specific apps can then make interesting things out of such content, e.g. create forms, save results... (there is an association in France which does that via boring copy-paste in PDF today).

If we have a highly structured format applicable to textbooks, usable in many countries around the world, some publishers will change their practices and create structured content from the start. Others will try to use AI to generate a proper structure. Whatever the workflow is, the result will be highly accessible textbooks.

Concentrating the work on structural semantics would be immensely helpful. Therefore we will be happy to participate to this discussion.

[AdamSobieski]

On the topic of educational exercises and activities, I would like to broach a new JavaScript API for bidirectional communication with services at schools. In these regards, interested contributors could reference: QTI/LTI/Caliper best practices.

In addition to structural semantics for highly-structured educational exercises and activities, we can also consider structural semantics for worked examples.

Use cases for structural semantics in digital textbooks also include Web-based and desktop-based search scenarios. Students could more easily search through the contents of their digital textbooks, course materials and notes with such features.

[redwidgeon]

Use cases for structural semantics in digital textbooks also include Web-based and desktop-based search scenarios. Students could more easily search through the contents of their digital textbooks, course materials and notes with such features.

I'm an educator, not a coder: apologies in advance for my lack of technical contribution. But it occurs to me that textbooks are not only [still] designed in very "print-first" ways but also with a very "read-first" emphasis. Textbooks are built on the old paradigm of read-memorize-regurgitate; read some more. Being able to search & add annotations & notes is a great start. In addition, textbooks (or associated apps) could add features strongly supported by learning science: interleaving, pacing, structured recall, & worked examples etc. which could be configured to control the degree of mastery required before the student progresses.

[AdamSobieski]

@redwidgeon, building on your ideas, we could support student modeling with granular skills or proficiencies, such as the various maneuvers of algebra.

Student model data, stored at schools, could be made available to students' digital textbook software through a JavaScript API. With such an API, client-side digital textbook software and/or server-side software at schools could better select and sequence content, exercises and activities for each individual student.

Interestingly, these ideas start to blur the lines between digital textbooks and intelligent tutoring systems.

[redwidgeon]

AdamSobieski, you said:

• Interestingly, these ideas start to blur the lines between digital textbooks and intelligent tutoring systems.

I think you are absolutely right. Textbooks are still a good resource once one has mastered the material. But in my experience they are a poor tool for learning, especially for a novice or print-challenged learner. A teaching resource that presented the material in chunks & included exercises to develop and assess mastery would (IMO) be far superior. The material, as it is mastered, could be added to a text resource that the student could use as a reference afterwards. Sort of a personal, "build it as you go" textbook.

[jkhurdan]

Hi, I work in the accessibility field for Rutgers University. Our staff here, and across the BigTen have been working with EPUBs heavily. We've also have been in close ties with Benetech, who certifies EPUBs for accessibility. I would love to help this group in whatever way I can.

[clapierre]

Thanks @jkhurdan we at Benetech are proud to be working with publishers to certify their EPUBs are Global Certified Accessible and @RachelComerford from Macmillan Learning is the first publisher to get GCA certified. We are very interested in this space and helping in any way to advance the user experience while ensuring what is created is accessible.

Ideally we would add these "progressive enhancements" in a way that those reading systems that can take advantage of these advance features can do so but allows any EPUB reading system the ability to access the book as well.

[science]

A group of OER publishers has been working on some ideas for structural semantics in K-12 textbook-scale resources for a little while. I'd love to find a way to contribute these approaches into this workstream. Our specification is here: https://github.com/K12OCX/k12ocx-specs -- also more human readable here: https://k12ocx.github.io/k12ocx-specs/

(We'd welcome all input). We have some components for lesson-level structure around sub-elements (we call "activities") that seems to be absent from the current spec? Also, I'm not sure if our concepts for "module" and "unit" level content organization can be remodeled in the Educational Structural Semantics draft or not..

I would welcome any next steps that we can take to try to merge in our concepts, where they are useful and valuable extensions to the current draft.

[bredd]

It seems to me that a key factor in a textbook or educational profile for EPUB would be the right metadata. Alignment to competencies or learning outcomes will facilitate locating the right material and reporting use (the Caliper / xAPI stuff that Adam mentioned). I'm a member of the LRMI working group (Learning Resource Metadata Initiative).

I haven't looked into the EPUB Metadata specs. But assuming they are based on Schema.org (which would make sense) then LRMI would be part of that.

[AdamSobieski]

In the specification which @science shared, we can observe the property of educationalAlignment which has one or more values of type AlignmentObject. We can observe the property of oer:hasLearningObjective which has one or more values of type oer:LearningObjective. We can observe the property of oer:doTask which has one or more values of type oer:Task.

The classes or types indicated (AlignmentObject, oer:LearningObjective and oer:Task) involve external models. I am a proponent of more granular models in these regards. In the domain of mathematics, for instance, models of interest would be sufficiently granular to describe specific problem-solving maneuvers and specific cognitive procedures or skills. More granular models would entail more granular student models.

As @bredd indicated,

Alignment to competencies or learning outcomes will facilitate locating the right material and reporting use.

On these topics, we might want to support collecting and reporting response time data, the duration of time from the start of an exercise or activity to its completion. In my opinion, the combination of more granular models and such response time data would provide educators, scientists and researchers with useful data with which to advance the state of the art with respect to topics including: psychometrics, mental chronometry, automatic item generation, automatic item sequencing, automatic item evaluation and item response theory.

[rickj]

So much has changed in the world of digital textbooks since EDUPUB was originally envisioned. I would not want to encumber future work based on those original, and now somewhat dated, assumptions.

redwidgeon's comment is just a part of the changes that are happening:

A teaching resource that presented the material in chunks & included exercises to develop and assess mastery would (IMO) be far superior.

[dalerrogers]

Hello all:

I am an ELearning developer and Assistant Professor of Multimedia Design at Nashville State Community College. I’ve been involved in the EPUB3 group for a little while. As someone who is investigating creating and using eBooks in my own curriculum, I see multiple areas where the technology could go.

Before I offer ideas, where can I find a copy of the existing documentation (no need to suggest something that already exists)? Does it reside in the gitHUB repository?

[ghost]

For general epub specification you might want to take a look at this https://www.w3.org/publishing/epub3/epub-overview.html

Today we just started the very first cg meeting. I think epub for education is definitely something very interesting to explore more. I wonder if anyone would be interested in form a task force in publishing cg to start collecting some use cases for digital reading x education.

[dalerrogers]

@rakutenjeff thanks for your prompt response. I know I missed today’s meeting. I’ll have to catch a future one.

Regarding being part of a task force, and collecting use cases, I wonder if The Daisy Consortium and the Alliance of Independent Publishers, can help in that regard? I’m following both groups. Both organizations are approachable and open with their information.

I live in Nashville TN, which is the corporate HQ of Rustici Software, the originators of XAPI. I realize before we get too complicated, book semantic markup would be the first goal. As Learning Stores become more widely accepted (the DoD funded the xAPI development), eventually if the technology is embedded in an ebook, consumers will be able to document the books they’ve read, and that information will be part of their prior learning history.

This would be classified as emerging learning technology. Is looking towards the future part of the mission of this group?

[iherman]

For general epub specification you might want to take a look at this https://www.w3.org/publishing/epub3/epub-overview.html

you might also want to follow the repository of the EPUB WG for the latest evolution.

[Jeffxz]

Let's try to get some use case in next week CG meeting.

[Jeffxz]

Have more members to discuss about education, wil l share later