Content must be accessible AND Reading System UI must be accessible
Websites for purchasing must be accessible (W3C - Web Content Accessibility Guidelines 2.0)
Semantically rich structured content is essential
Intelligent reading Systems use the semantics

4. Techniques

Structure the content using semantics
Identify content based on what it is, e.g. a paragraph is not a heading!
Order the content for correct reading order, i.e. the way it appears in a DOM
Images, flow charts, etc. require short and long descriptions
Video must support captioning and descriptive audio channel
Interactivity must be accessible, which is a big question

5. Key Benefits of incorporating accessibility

Legislative compliance.
Corporate Social Responsibility.
Highly sophisticated navigation scenarios can be created for semantically rich content i.e. gestures or keystrokes to jump to different landmarks in the book e.g. headings, footnotes, long descriptions ...
Highly sophisticated user interfaces can be created for semantically rich content on different devices e.g. an epub reader on a mobile phone that displays the long description for an image in a layered panel when the user moves her finger over the image.

Speech Synthesizer support

Markus Gylling

An increasing number of reading systems ship with built-in TTS
For the print disabled in particular, this is a crucial feature
But good (read: low rate of mispronunciation, near-correct prosody (stress, intonation) is a general reading system usability concern

Problems

The use of TTS within the DAISY realm over the past decade has identified the following core problems:

While TTS quality has improved, the quality varies dramatically between languages (the bigger the language, the larger the market for TTS vendors)
For academic and specific-domain content, most TTS's fail miserably (too many terms not covered by hardwired lexicon). For professionals and students, this means that consuming content using TTS is painful at best and impossible at worst.

Requirements

A way to link to publication and/or document-wide pronunciation lexica
A way to specify atomic pronunciation and prosody rules inline (lexicon-provided generalizations are not sufficient)

Solutions

Existing languages that we can reuse:

SSML

PLS defines a lexicon document type for pronunciation instructions
SSML 1.1 defines a set of elements that can be adapted for inline usage. (see also Z3986-AI which defines an attribute-axis-only adaption for SSML-based pronunciation instructions).

... note: SSML is partially or fully supported by TTS engines such as cepstral, AT&T, Loquendo, neospeech, and more.

CSS3

http://www.w3.org/TR/css3-speech/

Metadata

Metadata in EPUB

Bill Kasdorf, Apex CoVantage

It’s not whether metadata belongs, but which metadata belongs
What is essential?
What is stable (ideally, permanent)?
What vocabularies are widely used?

Standards are only part of the story

Most standards are created for specific purposes
Often have aspects useful to ePub
Usually have way more than we’d want to include
Usage is incomplete and inconsistent

Examples

MARC records—libraries
Dublin Core—universal, flexible, incl. non-text content
PRISM—publication (esp. magazine) oriented
XMP, IPTC, Exif, PSIR, etc.—image metadata formats
ONIX—bookselling

My focus today: ONIX

A communication or messaging format
Focused on selling books (print, eBooks, audio, etc.)
Very rich model—publishers use only a fraction of it
Mainly a marketing function
Constantly updated by publishers—periodic feeds

ONIX 3.0 (4/2009): Modular, enhanced for e-books

Message header
Record & product identity (13 identifier types)
Descriptive detail

<ProductFormDetail>EPUB</ProductFormDetail>
<EpubTechnicalProtection> (none; DRM; digital watermark; Adobe DRM)
<EpubUsageConstraint> types & limits (e.g., preview, print, copy/paste, share, text-to-speech)
Collection/part info
Title types and detail
Authorship/contributor info
Language
Extents, illustration info
Subject info (keywords, BISAC, BIC, LC, Dewey, MeSH, etc. (many international schemes))
Audience info

Collateral detail (descriptions, cited content, links, prizes)
Content detail (incl. page numbering, etc.)
Publishing detail (incl. © info, territorial rights)
Related material
Product supply info (markets, distributors, prices, etc.)

Organisations we should collaborate with on this

DAISY
BISG
BIC
Editeur
NISO

Issues with Magazines and Serial Publications

Dianne Kennedy, Peter Meirs

A Note

IDEAlliance PRISM has recommended XML metadata for publisher/aggregator interchange of articles
IDEAlliance is currently working to update XML metadata for delivery of article/magazines to eReaders
Working in coordination with IDPF

Issues

Title (is this the magazine title? The article title? How can we differentiate?)
In addition to publisher, Magazines need to identify the “corporateEntity” of the publisher
We also need “productCode”and “IssueIdentifier” in addition to dc:identifier unless these can be handled with opf:scheme
Dates: Magazines require a number of dates:

prism:coverDate
prism:coverDisplayDate
prism:publicationDate

Issues Article Identification Metadata

Identifiers for an article

prism:volume
prism:number
prism:issueName
prism:edition
prism:versionIdentifier
prism:startingPage
prism:endingPage
prism:pageRange

Issue: Article Subject Metadata

Article subject metadata

dc:subject
prism:keyword
dc:description
prism:event
prism:industry
prism:location
prism:object
prism:organization
prism:person
prism:timePeriod
prism:ticker
prism:genre

Recommendation

Develop mechanism to include PRISM metadata for serial publications within ePub We prefer to have ePub develop an extension mechanism to enable inclusion of PRISM metadata set for e-Readers than for IDEAlliance to develop a competing standard

Semantics

Dictionaries

Jeff Alexander, Intangible Press

Problems

EPUB is currently designed for books that are read from beginning to end, rather than for books that are used for very fast information lookup:

Dictionaries
Encyclopedias
Medical books such as drug information or diagnostic guides (ie, for looking up a drug by generic name "acetaminophen" or brand name "Tylenol", or a disease name in an alphabetical listing)
Possibly other types of books such as cookbooks, travel guides, etc.

Standard "find"-type search features are not sufficient for dictionary lookup
Two main use cases for dictionaries: 1) lookup while reading another book, 2) free-form lookup
Use case for glossary: lookup that is tied to a particular book
Dictionaries are complicated
Dictionaries contain a much wider range of special characters than other books
Dictionary data sets range from large to enormous

Requirements

A way to associate one or more search terms with a particular content region (a "dictionary entry")
A way to associate several types of searchable words with a dictionary entry:

Alternate spellings (eg, colour, color)
Derived forms (eg, "accessorize" at the end of an "accessory" entry)
Inflected forms which need to be searched but not displayed (eg, "runs, running, ran" for English "to run", or dozens of forms for verbs in French, Spanish, etc.)
Single word represented in multiple writing systems (eg, a Japanese dictionary gives both katakana and kanji forms)

A way for reading system to identify an ebook as a dictionary (eg, from metadata)
Accommodate both bilingual and monolingual dictionaries

Solution Guidelines

Performance! Must enable fast lookup, and it should take approximately the same time to look up "aardvark" and "zoo"
Flexibility to handle a variety of lookup-oriented content
Adaptable to a range of languages and to non-Latin writing systems

Book semantics in XHTML and the future of DTBook

Markus Gylling, DAISY

EPUB 2.0: DAISY XML (DTBook) as a text content format, alternative to XHTML 1.1
Ongoing revision to the DAISY standard (ANSI/NISO Z39.86-2011): distinct format-level separation between authoring and distribution

for authoring purposes, using a dedicated XML grammar (public draft at http://www.daisy.org/z3986/2011/)
for distribution purposes, intending to use a derivative of XHTML

Migration rationale

The cost-benefit ratio of using a non-XHTML-family grammar for distribution:

challenging to make use of web browsers as rendering components (increases cost and negatively impacts user experience)
AT created for the web not tuned to adapting behavior to elements in foreign namespaces
Semantics of XHTML are pretty close to DTBook anyway; in the minimal case, only a few extensions needed

Migration success criteria

Accessibility- and content integrity-wise, the following are success criteria for moving from DTBook to XHTML:

(accessibility of rich media and interactivity left aside in this talk)
ability to express content model restrictions (maintaining strict(er) requirements on document structure)
ability to extend the semantics (see below)

... think of the idea as a "strict profile" of EPUB XHTML with a few extensions, suitable for textbook domain content.

Building the list of extension needs

Remember: rich and predictable semantics is a fundamental requirement for building an intelligent reading system (the way DAISY defines "intelligent")

noting that (X)HTML5 is a better starting point than XHTML1.x given its (likely) inclusion of elements such as section, aside, hgroup, figure, math, ruby
general textbook domain elements identified as missing include:

notes and note references
annotations (original marginalia) and annotation references
equivalent of docbook:bridgehead (non-structurally-significant heading)
glossaries/dictionaries (in the backmatter sense, cf. #Dictionaries)
image groups (and rich (read:marked up) textual descriptions of images)
pagebreaks (authoring-time concern mainly)

... but of course, the list can be extended depending on whether drama, poetry and other types of content are included.

Solution options

Note that CDI (namespace-based extension) isn't the only alternative; attribute-axis semantic inflections can be used as well in some cases.

Consequences for EPUB

A win-win situation: instead of having to support two distinctly separate grammars, we move towards a situation where namespaces and (fundamental) markup models are shared. This reduces cost of implementation for reading systems/authoring tools/AT.

PRISM and Articles

Dianne Kennedy, Peter Meirs

A Note

IDEAlliance PRISM has recommended XML metadata for publisher/aggregator interchange of articles
IDEAlliance is currently working to modify XML metadata for delivery of article/magazines to eReaders
Working in coordination with IDPF

ePub Content Coding

Both PRISM/XML and ePub are based on XHTML
Coding of current book content is not as “rich” as magazines require

Navigation/findability
Layout

PRISM/XML handles this through the addition of prism:class= attribute

PRISM Class Attribute

body
box
byline
caption
credit
dateline
deck
footnotes
lead-in
pullQuote
sidebar
teaser

InLine Content Coding

Magazines have a requirement for inline content encoding to facilitate precise search

pim:event
pim:industry
pim:location
pim:object
pim:organization
pim:person
pim:quote
pim:keyword
pim:ticker
pim:timePeriod

Coding for Rich Media

Object is too simplistic to handle media
This revision of ePub should handle that
prism:media may be considered as a model

includes dc:creator, dc:contributor, dc:identifier for media object
includes prism:mediaTitle, prism:caption, prism:credit

Ad Material

We have a requirement to link Ad material to editorial content

Recommendations

Recommend a mechanism to support encoding of magazine/serial publication content within ePub

Navigation

Enhancements in the next generation NCX

Markus Gylling and George Kerscher, DAISY

As became clear in the 2.01/maintenance WG,

NCX was insufficiently documented in the specification (direct evidence that the intent of navMap and the availability of navLists were not made clear enough)
The direct lift-over of the DTD from the DAISY spec contributed to the confusion (multimodality, head metadata, @playorder that did not apply to the EPUB usage scenario)

ANSI/NISO Z39.86-2011 introduces the following major changes to the NCX grammar:

modularized, allowing creation of tailor-made NCX grammars for particular usage domains (such as EPUB)
allow expressions of machine-readable metadata for groups of, and individual, navigation items
ability to use (XHTML) markup in labels (i18n requirement)
increased terseness (smaller document size)

Need to collect further requirements from EPUB:

Examples from EPUB 2.0:

Multiple pagelists for the same publication (editions, and or paperback/hardback) (relates to machine-readable metadata)
Multilinguality/multimodality (are supported by the NCX in general - relevant in EPUB?)

Examples from EPUB 2.1/3.0:

Do we want to use the NCX to navigate rich media (say, a long video clip with distinct scenes/events) in addition to global document navigation?

Layout and Styling

Targetting Multiple Display Surfaces

Brady Duga, eBook Technologies

Traditional Layout

Layout to a single display surface (e.g. a page of known size) allows for certain assumptions that can determine a choice of font, margins, columns, etc. When new surfaces are thrown into the mix (e.g. a smartphone with 2 orientations, a tablet device, a computer with resizeable windows on a 30in monitor, text-to-speech) these assumptions fail.

Current State of Affairs

No conditional styling in ePub, with limited exceptions in CSS like em units and percent values
Display decisions on different surfaces left to Reading System which causes non-uniform content degradation
Best results obtained with minimal styles
Third party work-arounds

Not Just our Problem

Same problem exists today on the Web. What are the solutions there?

Server-based device detection (user-agent strings and such)
Conditional styling via media-queries (in CSS and html)
Dynamic styling using various scripting mechanisms

Option 1 is not available to us (no server in the mix), option 2 is not specifically supported in ePub and option 3 is not recommended (all reading systems are required to do is not display scripts to the user).

ePub is not the Web

Users and content creators expect more from their publications. Users will not be willing to pinch-zoom every page to a reasonable size.

Professional styling even in simple fiction
House styles preserved
Custom crafted content
Complex layout (textbooks, etc)
Styles preserved at various font sizes

Basic Requirements

Changes based on physical display size, ppi, color

Alter text characteristics (size, drop caps, etc)
Modify box properties (margins, padding, etc.)
Change flow (float to inline, number of columns, etc)
Adjust media properties (images, etc)

Also account for font changes by the user

Examples

(coming soon to a Wiki near you)

Core Requirements

A few core requirements emerge when we consider these features:

Manageable content creation

Styling for multiple sizes must be possible with reasonable resources

Implementable

Actual Reading System support is a must

From PDF to EPUB: Lost in Translation

Thomas Hadfield, CourseSmart

PowerPoint slides: From_PDF_to_ePub_Final.ppt

CourseSmart is a service provider selling Higher Education eTextbooks to students and providing electronic desk copies to instructors.

Our product delivery is primarily web-based, although we do provide a downloadable option via a partnership.

CourseSmart’s product was designed to solve several problems:

Lowering the cost of textbooks to students
Providing a lower-cost, environmentally-friendly mechanism for providing instructor evaluation materials
Utilize a very low-cost, easy workflow content ingestion process that lowers the barrier to publishers submitting textbooks
Provide a highly reliable DRM solution to protect our member publishers
Realizing that low-cost eTextbooks would co-exist with print textbooks for an indeterminate future, providing a format that allows side by side use of print and eTextbook formats

CourseSmart uses a “Page Fidelity” view providing identical pagination and formatting with the print book. The source files are the compositer produced PDFs that are converted to encrypted and watermarked JPGs with indexed text to permit searching and highlighting.

While this solution meets all of our initial product objectives, it is not amenable to:

Smaller devices that need reflowability
Requires more effort to support accessibility

Additionally, this format requires very large file sizes for large books.

CourseSmart has evaluated moving to an ePub standard, but has found the technology less than ideal as a solution to our problem set which we feel requires a rich layout. We have evaluated solutions such as HTML5 but are concerned that once text is available in the browser, our DRM is less effective.

Layout Requirements in Complex Documents

Ric Wright, Engineering Manager – Digital Publishing, Adobe Systems

PDF paper: IDPF_F2F_LayoutReqs.pdf

Interactivity and Rich Media

EPUB 2.1 Rich Media Thoughts

Garth Conboy, eBook Technologies, Inc.

PowerPoint slides: EPUB_2.1_Rich_Content.ppt

Demo EPUB (potentially evil, as it depends on Reading System knowledge; a direction we want to avoid in EPUB 2.1): SIMPLERICHCONTENT.epub

Rich Media and Interactivity - The Challenges

Ric Wright, Adobe

Problem

many speakers have mentioned the need for both animation and interactivity
iPad frenzy (soon to be Android tablet frenzy too)
applicable to all forms of media (books, newsreaders, magazines)

Requirements

Need a way to author both animation and interactivity
Books should work across platforms
Books should not break in (relative) perpetuity
Leverages known standards

Problems

Markup

No standard in sight

Although partially "standard", not perceived as such

HTML5/CSS3

"Publication as a Working Draft does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress. "

SVG Animation

1.1 is a recommendation, but now almost deprecated. SVG 1.2 was abandoned.

Effects of changes over time in standards
Name changes per se
Changes of underlying behaviour

Implementation

Performance

Range of EPUB devices from slow ARM processors with slow memory to fast desktops with GPUs
Slower devices may not handle interactivity little or not at all. Fallbacks?
Lots of interactivity may eat batteries

Scripting

Not all scripting engines are the same. What's the standard?
What can be scripted? A "core media type"? The actual text of the book (yikes)?

Display Technologies

eInk devices which are widely proliferated cannot support much interactivity or rich media

End-Game

Bottom line:

Are publishers willing to step up and request this support in the spec and acknowledge that this will result in books that degrade over time and potentially break completely?

Does the IDPF wish to bless a spec (HTML5/CSS3) which is encapsulated in the latest build of WebKit?

DAISY Multimedia Synchronization

Marisa DeMeglio, DAISY Consortium

What and Why

DAISY books are accessible eBooks which feature synchronized audio and text.

Some user groups for these types of books:

Print-disabled: dyslexic, blind, low-vision (DAISY users)
Mainstream readers of audio books with text.

How does DAISY do it?

Key components: structured text and human-narrated audio
Use SMIL to define synchronization points between text chunks with audio chunks
DAISY players contain SMIL-playing components which control audio playback and text highlighting.
This way, SMIL works seamlessly with the text and navigation. E.g., the user selects a chapter, and the book audio starts at that point. Or, the user has certain preferences, e.g. don't read page numbers (they get in the way of the narrative) -- the SMIL player skips over the page number announcements.

How could EPUB do it?

Already has structured text
Use SMIL to synchronize audio content
SMIL would be a transparent overlay on top of the document
SMIL could also be used to supplement islands of content within the document, e.g. SVG or Math
It is possible to use a limited subset of SMIL with low overhead
Reading devices that don't understand SMIL can ignore the overlay and not be affected by it

More information

See DAISY in action: AMIS DAISY Player video
See a visual for the proposed SmilOverlayFileset

Extensibility

Ben Trafford

PDF paper: IDPFf2fPaper.pdf

Extensibility in ePub

Why it's important

History

How we got here

Roadmap

What did we do before?

OEBPS 1.0 - 1.2

ePub 2.0

Future Extensibility

Supporting pure XML
Handling display behaviors
Cross-platform compatibility

Extensibility and the Charter

What can be rolled into extensions, and why

A Radical Suggestion

Why have preferred vocabularies at all?

Annotations

EPUB 2.1 Annotations

Will Manis, Viscous Documents

PowerPoint slides: Annotations.ppt

Advertising

EPUB 2.1 Ads

Will Manis, Viscous Documents

PowerPoint slides: Advertising.ppt

► Sign in to add a comment