understanding critical elements of e-books: the social reading experience of sharing bookmarks and...
DESCRIPTION
http://www.niso.org/news/events/2012/nisowebinars/ebooks_social_reading/TRANSCRIPT
Understanding Critical Elements of E-books: The Social Reading Experience of
Sharing Bookmarks and Annotations
September 12, 2012
Speakers: Todd Carpenter, Rob Sanderson, Dan Whaley
http://www.niso.org/news/events/2012/nisowebinars/ebooks_social_reading/
• INSERT EACH OF THE SPEAKER PRESENTATIONS, IN THE ORDER THEY ARE LISTED ON THE WEBSITE
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
29
Annotating the Web with W3C Open Annotations
Robert Sanderson [email protected] [email protected] Los Alamos National Laboratory @azaroth42 (W3C Open Annotation Co-Chair)
http://www.openannotation.org/
http://www.w3c.org/community/openannotation
This research was funded, in part, by the Andrew W. Mellon Foundation.
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
30
Open Annotation
• Communities• Annotation, E-Books and the Web
• Core Open Annotation Model• Extended Model: Specific Resources
• The Road Ahead
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
31
Open Annotation and the W3C
• Open Annotation Collaboration (focus: humanities, web)• LANL, UIUC, Queensland, MITH, plus others
• Annotation Ontology (focus: science)• Harvard, Manchester, plus others
• Models merged November 2011-March 2012• Basis of W3C Open Annotation Community Group• Next f2f meeting: September 18-19, Chicago • http://www.w3.org/community/openannotation
• Open to all• Please join in the discussion!
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
32
Social, Digital Annotation
• NISO Working Group on Digital Annotation• Apple, Adobe, IDPF, EBSCO, Sony, hypothes.is, …
• 3 Meetings (New York, San Franciso, Frankfurt)
• Introduced and reinforced several requirements:• Need to be able to specify how to render the annotation• Issues with text quoting and book reconstruction• Challenges with text segments:
• Robustness in the face of change• Precision of the selection
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
33
Further Community
• International Image Interoperability Framework• Stanford, Oxford, Cornell, BL, BNF, KB,
Natl. Library of Norway, UK Natl. Archives, ArtStor, …
• Digital Medieval Manuscript Interoperability Forum• As IIIF, plus Drew, St Louis, Ghent, Meertens Institute, …
• Close Ties to:• W3C Provenance WG• schema.org• NLP Interchange Format group
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
34
E-Books and the Web
• Web formats are E-Book formats• Epub is just HTML5• Apps/Sites and E-Books are moving closer together
• E-Books are not just Text!• Video or audio comments• … about embedded images, 3-d models
• Multiple targets for a single annotation• Compare/Contrast, grouping, …
• Fine grained region of interest is crucial• Particular word in text, letter depicted in image …
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
35
Basic Model
An RDF Document
The Comment
What The Comment is
About
http://www.openannotation.org/spec/core/
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
36
Basic Model
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
37
Annotation Types
Class Description
Bookmark Marker at (a point in) a resource
Change Request for modification
Classification Assignment of a class
Comment Commentary or Review
Description Description of, rather than about target
Highlight Highlighted (section of a) resource
Link Relationship of unspecified semantics
Moderation Assignment of value or quality
Question Question about target
Reference Citation or reference pointer for target
Reply Response to previous statement
Tag A tag on target resource, often textual
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
38
Annotation Types
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
39
Multiple Targets
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
40
Further Specification of Resources
Specific Body and Specific Target resources identify the region of interest, and/or the state of the resource.
Need to be able to describe the state of the resource, the segment of interest, and potentially styling hints for how to render it.
We introduce two Specifiers:
Selector Describes how to select segment
State Describes how to retrieve representation
(Where did Style go? It’s moving to the Annotation in next draft)
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
41
Specific Target
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
42
Offset Text Selector
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
43
Quotation Text Selector
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
44
Fragment Selector
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
45
Further Features
• Embedding resources • Embedding body, specifiers, styles for transport
• Semantic and Data Annotations• Semantic tags, annotating data, data as commentary
• Provenance of Annotation• Versioning• Equivalence• Archiving• Annotator / Generator distinction
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
46
The Road Ahead
• September face to face• Resolve open questions
• October-December• Revised Draft
• 2013• Final Draft• Look at moving to W3C Working Group structure• Towards standards track / recommendation status
Social Reading Experience of Sharing Bookmarks and AnnotationsSeptember 12th 2012, NISO Webinar
47
Thank You
Robert Sanderson [email protected] [email protected] @azaroth42
Web: http://www.openannotation.org/ http://www.w3.org/community/openannotation/ http://www.shared-canvas.org/
Slides: http://www.slideshare.net/azaroth42/niso-annotation-webinar
Overview
NISO Webinar on Bookmarking and Annotation
September 13, 2012
http://bit.ly/niso-hypothesis
”
Gutenberg invented the [modern] printing press, and so what was printed could now be distributed and affect otherwise docile citizens or subjects. It therefore became important to regulate that which was set before the public.
“Spier 2002 – The history of the peer review process
1913 Victoria Carbon Copy Machine
1960 Haloid Xerox 914
You are here
To transform how we know what’s credible in the world around us by
enabling the crowdsourced peer-review
of information everywhere.
Our Mission
Why don’t we have this yet?
http://bit.ly/existingannotation
Why don’t we have this yet?• No effective peer-review model, if
any.• Not annotation based—i.e. many
don’t work on subparts of documents.
• No way to link into passages and annotations.
• Poor cold start strategies.• Not standards based.• Not open source.• Not non-profit in structure.
What is Hypothes.is?
1. Open source. Free. (github.com/hypothesis)– Leverages OKFN annotator work
2. Non-profit.3. Enabling broad annotation of the world’s
documents and other media.4. Using a pseudonymous, occasionally
anonymous identity model.5. With an emergent & domain sensitive
reputation model.6. Coordinating with others to achieve long-term
archiving.7. Committed to interoperability and data
portability
FY 2012-2013 Funding
$245k*
*includes matching gift and others
$520k +$20k workshop
A few Grand Challenges
1. Effective reputation model (signal v noise) that makes efficient use of experts (metamoderation).
2. Robust inter and intra-document anchors
3. Works well across the web and with a diversity of browsers, tools, etc.
4. Great UI5. Solving the cold start
Timeline
2012 2013
Annotation / Anchoring / UI
Cold start
Reputation model / Peer selection
now
Scale
Annotation Peer-review
+
Annotation
Anchors• Inter-document
– How to create robust anchors that survive document changes in structure and content
– Strategies for calculating confidence metrics and graceful failure modes
• Intra-document1. Different document formats (HTML / PDF / Text)
should show same annotations (related to Canonical URL issue)
2. Handling pagination vs. single page (related to reflowable text)
3. Pure content anchoring (annotating Hamlet, or the King James bible wherever it is)?
Now
Blog “B”Document “A”
Coming soon
Blog “B”Document “A”
Blog “B”Document “A”
Blog “B”
17m 32s
Blog “B”
Blog “B”Document “A”
ChallengeSupportNeutral commentPose a questionAnswer a question
Request expert reviewSupply a correctionJokeProvide a referencePoint to similar articles
Document “A”
Document “A”
Peer-review
It doesn’t have to be complicated
“I just say something, and they say ‘yeah, but have you considered X?’
And I say ‘Huh. Good point.’ ”
Ted Underwood / UIUC – Dept of English commenting on the objectives of a good open peer-review model
The Reputation Modeling Workshop
• Held Feb 22-24, 2012 in San Francisco
• Funded by $20K grant from the Sloan Foundation’s Digital Information Technology area
• Purpose: To bring together leading experts in reputation modeling to think through design challenges of enabling global annotation
• Over 50 attendees including:– Erik Martin / Reddit, GM– Jeff Atwood / Stackoverflow, Cofounder– Rob Sanderson / LANL– Radu Jurca / Google (Zurich)– Steven Tadelis / eBay– Dario Taborelli / Wikipedia– Roger MacDonald / Internet Archive– Paul Resnick / U Michigan– Jennifer Lin / PLoS– Craig Newmark / Craigslist
What we learned
• The identity model, and particularly how expensive new identities are, is central to the dynamics of the reputation model.
• We need to balance direct participation w/ an objective control that should be semi-random and probably blind.
• It needs to be fun.
Signal to Noise
Obvious trolls
Identity
Voice
SQUELCH
Noise floor
Meta-moderation
Spam
Signal
Noise
VOLUME
Design choices
• Pseudonymous identity, potentially gated through two factor authentication (i.e. SMS message).
• Threaded, direct reply approach to moderation where sentiment and score propagate up through threads
• A meta-moderated loop controls reputation• Meta-moderation is blind and semi-random• Meta-moderators are chosen by their
reputation in the associated area. By domain proximity.
• Users can annotate for personal use, in either public or private groups, or in a global public channel.
• Standard social kit– following others, sharing, etc.
Two moderation loopsModeration Meta-moderation
Opportunities Come through replies in threads
Come via an activity stream
Participants Involves people that are already at the site
Blends people between sites based on expertise
Reputation … Is leveraged Is determined
Interaction Model
Direct response in threads
Blind, random
Qualifications Anyone who has passed initial screen
Only people above a certain reputation threshold
Economics
Monetary Policy
Oceanography
Climate Modeling
Human Rights
Quantum Teleportation
LHCPhysics
Expert matching through domain proximity
Economics
Monetary Policy
Oceanography
Climate Modeling
Human Rights
Quantum Teleportation
LHCPhysics
Meta-moderation matching through domain proximity
Economics
Monetary Policy
Oceanography
Climate Modeling
Human Rights
Quantum Teleportation
LHCPhysics
Meta-moderation matching through domain proximity
Plug-ins
URL
Deep links / embeds
Website
API
A diversity of interfaces
http://hpt.is/TzLv
Internet Archive partnership
• Will store annotations permanently• Domain assignment agreement• Will cache reference copies of annotated
pages permanently (for versioning)– Exploring same relationship w/ Common Crawl
• Wants to enable annotation on their – Wayback machine (version history of internet)– Books– New video project
– Early, active & enthusiastic OA participant
Use case examples + formats• Gov’t documents
– Legislation, regulations, FOMC minutes
• Scientific papers– Post-publication review
• Terms of service, Privacy policies, EULAs– Facebook, Yahoo, OnStar, etc
• News / Blogs• Data (scientific, public, industrial,
etc)• Corporate documents
– 10K, 10Q, analyst call minutes
• General content
• HTML• PDF• EPUB• Video• Audio• Images• Manuscripts• Data• Software• Spreadsheets & other
application-specific document formats
Selection of launch domains*• Stable (easy to anchor to)• Longevity (at least months)• Single sources (canonical URLs)• In the public interest • Trending• Content is not highly subjective• No commenting system already• Large pool of quality annotators• No paywalls• Annotation makes a difference
* aka Solving for the Cold Start
Use case examples
• Gov’t documents– Legislation, regulations, FOMC minutes
• Scientific papers– Post-publication review (arxiv, PLoS)
• Terms of service, Privacy policies, EULAs– Facebook, Yahoo, OnStar, etc
• News / Blogs• Corporate documents
– 10K, 10Q, analyst call minutes
• General content– Books, PDFs, video, etc
Thanks!