15 January 2013
When designing and developing both IGP:Digital Publisher and AZARDI we have put a lot of effort into using the presentation and behaviour semantics available with the epub:type property.
When designing and developing both IGP:Digital Publisher and AZARDI we have put a lot of effort into using the presentation and behaviour semantics available with the epub:type properties atttribute.
For those who don't know about this, it is an attribute that can be added to an appropriate element to identify the elements structural or semantic value and possibly purpose for an aware reading system. You can see the IDPF Structural Semantic Vocabulary specification here.
It is very labour intensive to apply these properties accurately and in detail in an ePub book unless the tools being used allow both custom attributes and values to be applied automatically; and/or the underlying XHTML is structured correctly. Most desktop tools such as Sigil, InDesign and others don't have provision for structural semantics on XHTML CSS selectors in a consistent manner.
Exploiting the epub:type properties was relatively easy with IGP:Digital Publisher because the underlying XTHML is IGP:FoundationXHTML (You can see the very large details of IGP:FoundationXHTML (FX) here).
FX has a very large and fully defined set of semantic and processing class selectors for a wide range of content including trade, academic, specialist genres (such as poetry, drama, textbooks, interactive content, QAA and more). It is easy to extend, maintain and customize. More importantly it is natively applied in IGP:Digital Publisher.
Because the descriptive vocabularies in standard print publishing (and standard digital book publishing) have been established for decades (and some for centuries), the FX vocabulary and IDPF epub:type vocabulary were a close match. Of course the IDPF vocabulary does not yet support properties for rich media, multimedia or interactive structures beyond Media Overlays.
Because we don't want to dirty our master FX content with anything as tacky as custom format properties these are processed in at format creation time (they are likely to change in the future or become irrelevant in this fast changing digital world).
During ePub3 packaging IGP:Digital Publisher has a cross-walk map between the IGP:FoundationXHTML structural and semantic class selectors and the IDPF epub:type properties. This is all handled in depth and automatically.
You can see this in any Infogrid Pacific produced fixed layout ePub3 by using the Alt-H key combination to explore the applied properties.
Basically nothing. They are there for reading systems. We are currently in a Catch-22 situation with "there are no readers that support epub:type therefore there is no reason to tag epub:type in ePub3 books". But if a reading system understands them they can be used to good effect.
There is no specification which states how these should be used, but the AZARDI ePub3 reader uses the following epub:type properties "experimentally" at present.
epub:type="cover" Centers the cover viewport if it has no other spine properties and it is a spread.
epub:type="copyright" Sets the viewport to the left if it has no other spine properties and it is a spread.
epub:type="titlepage" Sets the viewport to the right if it has no other spine properties and it is a spread.
epub:type="part" Sets the viewport to the right if it has no other spine properties and it is a spread.
epub:type="pagebreak" AZARDI has a Page button to hide or show page numbers as inline, in the margin or as a bar. This was seen as significant and important, especially for academic and education books. In mixed e-Book/print instructional environments the "turn to page 286" can be easily carried out by all.
There is no other extra use of the epub:type properties in AZARDI for reflowable books.
Apple did one good thing in iBooks and enabled pop-up footnotes using the epub:type="noteref" associated with the epub:type="footnote". This is described in detail by Liz Castro here.
Unfortunately it does not handle epub:type="endnotes", probably because most production systems don't have the capacity to move book endnotes into the same file to allow these actions and they are a lot denser in books that use endnotes.
AZARDI supports both footnote and endnote pop-ups, but the endnotes don't show in iPad. To make that work we would have to cheat and make endnotes into footnotes and that is dangerous territory moving forward. Reminiscent of tagging compromises inflicted on content because of the limitations of Mobi and ADE in the past.
We also introduced a non-standard property for our own customer use and that was epub:type="glossaryref" so we can automatically bring up glossary definitions in education books. That was probably a little naughty, but we needed it and having one consistent method of creating and running popup text made sense.
One of the most difficult, but valuable pop-ups is indexes. Especially so for academic books. Of course very few ePubs have index reference that are linked directly to index terms. IGP:Digital Publisher can be used this way and it generates both print and e-book indexes. The IDPF vocabulary doesn't support both ends of indexes at present either.
However this is now practical to consider because ePub3 acknowledges correlation between print and e-book page numbers. That means we don't have to hide page numbers with silly substitute link systems anymore.
We are currently working on the processor to allow an on page Index item (reference) to link to its index term. That will of course pop-up with a number of other associated index term links. It will be possible to traverse associated links serially and land back at the exploration starting point. Rather cool (I hope).
Posted by Richard Pipe