Adobe EpubCheck to validate Epub documents

[Alexander Turcic]

Adobe EpubCheck 0.9.1 is a Java-based command-line tool to validate IDF Epub files. It can be used to detect various types of errors in an Epub OCF container structure and in OPF and OPS mark-up, and to check for internal reference consistency. While this tool is probably not for your average joe, early adopter Epub geeks may find some use for it.

Now this tool (named EpubCheck) is available as an Open Source project. It is not complete (there are still many checks that we can do), but it is already fairly mature and extremely handy. If you author epub files, you should consider running this tool on your content regularly.

[via Adobe Digital Editions blog]

[NatCh]

It's not the mobile DE we're drooling for, but it's still good to see signs of progress on the DE front.

[TadW]

Good news indeed. I see Epub as the competitor with the most potential to compete against Amazon's ever growing (e-)book empire.

[NatCh]

I see epub as the best hope for an end to the tower of e-babel. Time will tell whether any of that comes to pass.

[GregS]

Forgive the cross posting but I wanted to repeat what I found here:

This is a critical tool.

Adobe's version of "Alice in Wonderland", and "The Adventures of Sherlock Holmes" both are full of errors.

Even "EpubGuide-hxa7241.epub" has the wrong mine type and missing attributes in the toc.ncx file.

It will help, at the very least, software developers to better write reader software, and it will be invaluable to publishers and editors.

Could I make an open request to anyone more HTML savvy than myself - can the jar file have a HTML front end to make it a bit more user friendly (with a file selector and a text file output)?

[DaleDe]

Originally Posted by GregS

Forgive the cross posting but I wanted to repeat what I found here:

This is a critical tool.

Adobe's version of "Alice in Wonderland", and "The Adventures of Sherlock Holmes" both are full of errors.

Even "EpubGuide-hxa7241.epub" has the wrong mine type and missing attributes in the toc.ncx file.

It will help, at the very least, software developers to better write reader software, and it will be invaluable to publishers and editors.

Could I make an open request to anyone more HTML savvy than myself - can the jar file have a HTML front end to make it a bit more user friendly (with a file selector and a text file output)?

Most, if not all, of the digital editions books were built before the spec was finalized. this is why you found errors. Actually all the books are likely to have errors.

Dale

[Hadrien]

Originally Posted by DaleDe

Most, if not all, of the digital editions books were built before the spec was finalized. this is why you found errors. Actually all the books are likely to have errors.

Dale

Yes, that's the main reason that we need such a tool. Now that it's available, I'll use it to improve Feedbooks too !

[GregS]

DaleDe I assumed that was the reason, it underscores the importance of the tool. It gives us an absolute measure - if it passes this test then the reader is at fault, and that helps for bug bashing and feature enhancement.

Hadrien I have just come across Feedbooks I am very impressed by what you are doing there.

I have some suggestions for your epub software.

I suggest the mandatory use of rational paragraph numbering for IDs (ie Chapter I.35) referencing is important in some circumstances and page numbers are dead for this in ebooks.

Use the XMLnamespace effectively by using World Unique Ids (WUIDS) and not urls.

If you are interested in either of these drop me a line at greg.schofield@iinet.net.au

PS WUIDs can be generated from any server with a simple script. Hence xgdgsgwf352sgsdesw36ff.I.35 is an absolute electronic reference. I am going to test out XMLnamespace WUIDs when I can to see whether this throws any errors through the testing app.

[hxa7241]

Originally Posted by GregS

Even "EpubGuide-hxa7241.epub" has the wrong mime type and missing attributes in the toc.ncx file.

That caught my attention, so I checked it...

There really doesn't seem anything wrong with my epub (it has the correct mimetype and conforms to the ncx DTD) -- so it seems my epub has found errors in the checker rather than the other way around.

(Thanks for the 'even'.)

[GregS]

Originally Posted by hxa7241

That caught my attention, so I checked it...

There really doesn't seem anything wrong with my epub (it has the correct mimetype and conforms to the ncx DTD) -- so it seems my epub has found errors in the checker rather than the other way around.

(Thanks for the 'even'.)

hxa7241 I am trying to work my way through things re epub. Your example/guide has proved most useful for many reasons, not the least because as a reliable example it easily shows the difference treatments between ADE 1.0, Mobireader 6.1 and the latest FBreader.

I was surprised to the extent that CSS descriptions were apparently ignored or parts not implemented by the various readers.

I would very much like to see other examples, by someone like yourself who knows what they doing. I am at only the begining of things, trying to work out a tools set to use and how to go about things sensibly. Planning in otherwords at this stage, of what should be aimed for and whether the aims are compatible to the standard.

I am very interested in using other than XHTML, DTBook, but read also that arbritary XML might be included, I am thinking of TEI. "Page-template.xpgt" I cannot make up my mind about it - whether it is useful generally or something that can be, or should be, dispensed with.

Like many people I have problems reading specifications and making solid and practical sense of them. A few examples therefore go a long way.

For instancde I would like to use the the ID attribute to give each ebook a reference system, for students it is important that the paragraph number be displayed but for ordinary reading this is something that is better hidden. I cannot work out if alternative stylesheets are provided for within the specs for epub (even if not implemented by any reader software).

Also, I do not know how tolerant Epub is of including foriegn material. I am working on (very slowly and early days yet) of a stylesheet language that allows for PDFs to be produced in a variety of forms based on the standard files of the Epud.

A forum devoted to Epub might be worthwhile, or a reference to one that already exists (I have not found one). In short, I have dozens of questions about what can, and cannot, be done with this standard.

I intend to produce ebooks that are well marked-up and reliable (non-DRMed) texts, with features (a robust reference system, library numbers, and unique identities for work, author, edition, publisher etc.,. and when micro-payements become available - a copyright disbursement system in digital signatures).

And things simple, like the ability to paste in an "Exlibris label" page after the cover if the customer wishes and a digital signature of receipt information.

This is an exciting period for ebooks, and I have occupied the shadows for many years looking on, waiting for eink/epaper technology and a standard to emerge. Now everything seems to be in place for this technology to take its proper place.

All the best.

[hxa7241]

GregS: I have another example -- a small book by Peacock -- on my books page http://www.hxa.name/books/. It includes a variety of bookish format features -- footnotes, poetry, various headings, small-caps, even greek text. It may be good for comparing readers.

On page-template.xpgt: I would consider it an isolated refinement for Adobe's reader(s). Other things -- authoring, packaging, etc. -- should not depend on it. It is minor. I have used it because the DigitalEditions defaults are a little weak in choosing column width and margin size.

On alternative stylesheets: epub explicitly mandates support.

On foreign content: epub explicitly does not mandate display of all/any unicode chars.

But we are dependent on reader-software/devices. And they will not all implement all the spec. For content projects, I, personally, would plan only around the core XHTML+CSS. Anything else has various amounts of risk.

[GregS]

hxa7241 you are a gem, this is just the type of examples I need to get up to speed. And in making comparisons it is invaluable - many thanks.

I must have read about alternative stylesheets somewhere on the IDPF site, but later when I wanted to make sure, I could find no reference. Your confirmation is as good as gold as far as I am concerned.

Your advice on using XHTML+CSS will be adhered to as well, at least to begin with, in the end hopefully all the readers will improve in this regard.

By the way your practical guide has been the best resource for clarifying the standard that I have come across, I am in your debt.

I hope, as things unfold, the standard embraces a number of “standard” DTDs along with arbitrary XML — DTbooks has virtues, TEI is essential for serious academic work (it really needs a collection of stylesheets to be anywhere near functional), DocBook etc.,.

The foreign matter I was thinking to include, is part of long term project, that is providing typographical stylesheets that produce PDF for readers to display (I don't expect the readers to do this transformation, rather to have the capacity inbuilt into the epub and produced on demand via a local computer).

Early in 2008 I hope to put in the public domain, with the help of Sydney based company, an easy and reliable system for creating world-wide unique identifiers. That is a hole that needs fixing sooner rather than later, along with a register of works, authors, editions and publishers based on unique identifiers.

Exciting times for e–books ahead, epub need not do everything, but as a standard it is sorely needed.

Many thanks for your answers and most of all your examples.