OPDS Standard Compliance

[Frank Lowney]

I've written a book entitled, "The Coming ePublishing Revolution in Higher Education" that I am currently revising. On the iBookstore, eBooks are versioned like software so this will be a free update to a 99 cent eBook. More on that here. Part of the book is dedicated to illustrating how academics could create and distribute their own teaching/learning content using the *.epub and *.ibooks formats. In this updated, expanded and revised version, I wanted to say more about OPDS as a potential distribution option with distinct advantages for educators. I make extensive use of  screencasting in the book so would use video plus text and images to illustrate this.

An OPDS catalog might be an ideal way to distribute self or colleague-created learning content to students, especially free *.ibooks files that need to be organized in ways that the iBookstore cannot accommodate. Toward that end, I have looked at Calibre Server, Calibre2OPDS and COPS Server.  My test bed Calibre server OPDS feed may be found here. Leads to other server options that a university professor might be able to field will be appreciated.

The issue that currently has me stymied has to do with my observation that OPDS servers and clients may not be referencing the same standard and, consequently, a Tower of Babel situation has arisen. For example, KyBook has this pass/fail list of public OPDS catalogs it tested. My own testing against private OPDS catalogs such as Calibre indicate a  similar pattern - some work, some don't. My sense is that some servers and clients adhere to the OPDS standard that Lexcycle developed for Stanza (now owned by Amazon and no longer in development or supported) while other servers and clients adhere to the official standard now in version 1.1 with version 1.2 in the offing. 

Compounding the problem, many OPDS clients are tied to a specific eReader. As fragmentation increases with the deployment of ePub 3 plus the increasing popularity of the *.ibooks format among academics, I anticipate increased frustration with OPDS as a discovery route. The eReader bundled with OPDS client will discover many eBooks that it cannot display well or at all. Thus, I have begun to look for standalone OPDS clients finding eBook Search for iOS and Vienna for MacOS X - the two platforms I can most easily test on. These apps enable browse/search of OPDS catalogs, downloading selections and even a hand-off to a user selected eReader. This would nicely deal with the fact that the iBooks app will probably never support OPDS.

While these two apps work wonderfully with Gutenberg and the Internet Archive, they work poorly against the private servers: Calibre, Calibre2OPDS and COPS.

Most academics already have a full plate and, so, adoption of OPDS  as a distribution  method needs to be much surer and simpler than this. I'd love to be able to present OPDS as a viable distribution option but the current state of standards compliance seems to militate against that.  I am posting all of this in the hope that folks here will help me see this differently.  Perhaps  there is a private server I've not  yet discovered, a server that a university professor could run on their office computer or a virtual server on campus.  That server might be manageable via FTP or WebDAV using folders to generate categories into which *.epun and *.ibooks documents are dropped.

[Hadrien Gardeur]

Hello Frank,

The current state of OPDS is like any other standard: unevenly distributed.

To make the ecosystem move forward, several things have been done in the past and continue to be actively improved, such as:

• a test catalog for OPDS clients: http://feedbooks.github.io/opds-test-catalog/
• full test results for some of the most popular clients are available at https://docs.google.com/spreadsheet/ccc?key=0AiWt3tA8qtyMdFpkZ0lxV011MXV6X3JJRFUxa050SHc&usp=sharing#gid=0
• a feed validator is available at http://opds-validator.appspot.com/

As you've correctly pointed out, some clients/catalogs are stuck in the pre-OPDS world (what we used to call Stanza catalogs) and others have structured their catalog without respecting the navigation/acquisition structures (that's the case of the Project Gutenberg catalog for example).

We could potentially publish test results for catalogs and catalog servers (such as Calibre, Calibre2OPDS and COPS) like we do for clients.

In my own experience, I know that Calibre2OPDS has done some good work in being OPDS 1.1 compliant, and usually Calibre does a bunch of incorrect/invalid things. If Calibre2OPDS allowed a folder as an input and provided an option for dynamic catalogs (supporting search and live updates), it would be exactly what you're looking for.

On our side, we try to provide a fully valid catalog using all of the most advanced OPDS features at Feedbooks. We've also acquired Aldiko which as you can see has the best OPDS support on the client side, and our goal for Aldiko is to reach full compliance in the near future.

For EPUB support, I hope that you're wrong about two things: I don't think that the *.ibooks format as a proprietary alternative to EPUB3 has a lot of future, and thanks to Readium SDK we'll see a lot more apps/devices supporting EPUB3 this year.

Hadrien

[Frank Lowney]

On Jun 16, 2014, at 11:12 AM, Hadrien Gardeur <hadrien...@feedbooks.com> wrote:

For EPUB support, I hope that you're wrong about two things: I don't think that the *.ibooks format as a proprietary alternative to EPUB3 has a lot of future, and thanks to Readium SDK we'll see a lot more apps/devices supporting EPUB3 this year.

It may well be that *.ibooks and *.epub version 3+ will eventually converge. The implementation of widgets is the only big difference.  However, the popularity of the *.ibooks format right now is due to the availability of iBooks Author, a free and very powerful eBook editor.  There is nothing like this in the ePub 3 world.  The closest competitor to iBooks Author, IMO, is Pages, another Apple product that's bundled for free with MacOS X 10.9 which is also free. Pages with its ePub export supports a few aspects of ePub 3. 

I do have high hopes for Readium.js as the next step up from what Ibis Reader demonstrated. For educators, having a single eReading environment that is free and fully cross-pltform by virtue of being web-based would be ideal. Content, though, is still king and, so, the power and ease of use of available editing environments will be powerful drivers. Perhaps there should be a Writium Project to complement Readium.

Hadrien

==================================================================

Dr. Frank Lowney  Georgia College & State University

    Projects Coordinator, Digital Innovation Group @ Georgia College

   Chappell Hall 212 (CBX 106)

    Web Site, Blog, GCSU Email, iCloud Email

    My latest book: The Coming ePublishing Revolution in Higher Education

Voice: (478) 445-1344

NOTICE: Please be advised that I am hearing impaired and communicate most effectively via e-mail.  Follow-up summaries of telephone conversations by e-mail are most appreciated.

==================================================================

We don't make instruction effective, we make effective instruction more accessible.

[Frank Lowney]

On Jun 16, 2014, at 11:12 AM, Hadrien Gardeur <hadrien...@feedbooks.com> wrote:

As you've correctly pointed out, some clients/catalogs are stuck in the pre-OPDS world (what we used to call Stanza catalogs) and others have structured their catalog without respecting the navigation/acquisition structures (that's the case of the Project Gutenberg catalog for example).

We could potentially publish test results for catalogs and catalog servers (such as Calibre, Calibre2OPDS and COPS) like we do for clients.

That might be helpful. Using http://opds-validator.appspot.com/ I find that both Calibre and Calibre2OPDS feeds are not valid. The COPS feed was valid.

Since it takes two to Tango, both clients and servers  need to be  talking the same language. As I mentioned, I am particularly interested in standalone OPDS clients 

Using the OPDS Test Catalog, I find:

eBook Search for iOS  (https://itunes.apple.com/us/app/ebook-search-free-books-for/id416454511?mt=8) works as expected

Vienna for MacOS X (http://www.vienna-rss.org) gets the toplevel listing but is unable to navigate to lower levels.

Would it be possible for me to contribute a test *.ibooks file to the test feed to see what happens?

[Frank Lowney]

Here's a  bit of related news. The latest version of the iOS app called "Book Search" has added  the ability to see *.ibooks files in an OPDS feed (I tested Calibre), download them and hand them off to the iBooks.app.  

This is a breakthrough for educators who are producing free eTextbooks for their students using iBooks Author and who also wish to present those free eTextbooks in ways that are not possible with the iBookstore and difficult to maintain on a regular web site.  This app seems to do well with COPS server as well although the catalog I tested (http://cops-demo.slucas.fr/feed.php) did not contain any *.ibooks files.

On Jun 16, 2014, at 11:12 AM, Hadrien Gardeur <hadrien...@feedbooks.com> wrote:

As you've correctly pointed out, some clients/catalogs are stuck in the pre-OPDS world (what we used to call Stanza catalogs) and others have structured their catalog without respecting the navigation/acquisition structures (that's the case of the Project Gutenberg catalog for example).

We could potentially publish test results for catalogs and catalog servers (such as Calibre, Calibre2OPDS and COPS) like we do for clients.

That might be helpful. Using http://opds-validator.appspot.com/ I find that both Calibre and Calibre2OPDS feeds are not valid. The COPS feed was valid.