The Return of the Prodigal Web Carole Goble Information Management Group The University of Manchester, UK [email protected]
Oct 17, 2014
The Return of the Prodigal Web
Carole GobleInformation Management Group
The University of Manchester, UK
2
The Return of the Prodigal Me
Semantic Web
E-Science(life science)
Distributed Computing
Hypermedia
Web Accessibility
Hypertext?
Web?
Hypertext?
4
What’s the Story?• Hypertext, and its offspring the Web, envisioned
collective intelligence.• Both propose technical infrastructure and social
infrastructure to do this.• The Web “made riotous living”.• Two evolutions of the Web – Web 2.0 and Semantic
Web – could (potentially) return the Web to the Hypertext family. By stealth.
• We have applications that show this.• “We should make merry, and be glad”.• How? Is it done? Lessons learned? Why bother?
5
Spot the Difference Competition
• The Return of the Prodigal Web, Carole Goble
• Back to the Future with Hypertext: A Tale of Two or Three Conferences, Wendy Hall
• Back to the Future: Hypertext the Way It Used To Be, Ted Nelson and Robert Adamson Smith
• Hypertext Tragedy, A Panel
Is Simon trying to tell us something?
Back in the Day
Or Back to the Future….
7
Xanadu: a system in which it will be possible to store anything that anybody has ever written (or photographed or filmed) that can connect any piece of information to any other piece of information
Augmentation of the human intellect.
“ Adequately equipped with machines which leave him free to use his primary attribute as a human being – the ability to think creatively and wisely, unencumbered by unworthy tasks – man can face an increasingly complex existence with hope, even with confidence.”
“the interaction between person and hypertext could be so intuitive that the machine-readable information space gave an accurate representation of the state of people's thoughts, interactions, and work patterns, then machine analysis could become a very powerful management tool, seeing patterns in our work and facilitating our working together through the typical problems which beset the management of large organizations.”
A universal medium for
harvesting and harnessing collective intelligence
making and using connections between
decoupled content and decoupled people.
9
How? Hypermedia
• Systems that present the same media (text) in multiple CONTEXTS
• Systems that enable the communication of RELATIONSHIPS between documents
Frank Shipman
10
Technical Infrastructure– Management of nodes and
links and anchors.– Combining and aggregating
content and structure.– Architectures and models for
these connections.
Social Infrastructure– Authoring.– Navigating and Finding.– Making connections between people
and content.– DIY content aggregation.– Adaptation. Reuse.
11
Build Hypermedia
Create Content and Connections
Cruise it. Use it.
12
It’s the LINK, baby• LINKs as first class citizens• Embedded and non-embedded LINKing
• LINK bases, open hypermedia
• LINK architectures and LINK models
• LINK typing, LINK properties
• Two way LINKing• LINK context, LINK filtering
• LINK trails, Computed LINKs• LINK management
• LINK behaviours, Dangling LINKs• LINK navigation, Adaptive LINKing
• Rhetoric of LINKS• LINK authoring• Reusing resources through different LINK sets.
• LINK foo. LINK bar.
14
Open ArchitecturesAnchor object
Link object
Anchor object
Node Object Node Object
Link Base
Link Service
My Favourite Application Your Favourite Application
17
Open Composite Models• High level
representation• Dynamic• Virtual• Reuse
• We would call this a mash-up aggregating content feeds today.
I’m sure Ted will tell you all about Transclusion.
19
Open Authoring• Reuse
– content for a different audience.
– adaptation of materials you don’t control
• Alternate paths• New paths• The annotation thang.• Fuzzy distinction
between authors and readers. Possibly.
Open
Well yes…but aClosed System mentality.
Come and look in MY system.Kind of assumption that content was limited and
controlled
Closed Authoring and Closed Architectures
21
The Web
22
TBL’s original LINK fest
1989
23
Simple Hypermedia• Simple.• Distributed.• Low barrier of entry• Network effect• Long tail
• Open Architecture• Robust. Don’t restrict,
fix! Cooperate, don’t control.
• Scalable.
24
Simple Hypermedia• Simple link types• Composite nodes• Virtualisation• Computed links
• All added in through rich browser and server technologies and applications.
• CMS
25
• Search • Composites • Virtual • Extensibility • Computation • Open Standards
• Versioning • Collaboration
Halasz’s Seven Issues
Browsers and AppsPortal Frameworks
CMS & EnterpriseServers
27
DLRScirca 1998
Publisherslinkbases
Peopledatabase
Keywordlinkbases
Proxy
Hyperlink
Agent
XLinklinkbases
User profiles
Person Agent
Citation Agent
Bibliographydatabase
IR AgentInverted
index
Ontology Agent
Web Browser
Web Server
Carr, Hughs, Hall, De Roure et al,
University of Southampton, UK
28
Do links appear? Sure!• Google rocks because of Links.
• XLink – but that never took off…..
• Open hypermedia in Content Management Systems on a big scale – Active Navigation - Dynamic Link Injection – Hyperwave - Collaborative Information Management”
• Web + Open Hypermedia is an extra layer of software engineering.
• Wikis
But lets face it, its still Node-CentricThere is no Link Model.
Makes Context and Management Harder.
Links are embedded in the Applications.
Captive in their systems.
No infrastructure outside the application or the browser.
No Third Party Link Authoring (except shared bookmarking)
User participation is Reader Participation.
30
The Parting of the Ways…
The Web “made riotous living”
33
Hypermedia on the Web.
Versioning. Collaboration.Links. No Link Model outside the wiki.
34Collective Intelligence
• O’Reilly Publishing 2005ish.• Bottom Up & Emergent.• A collection of in use memes,
patterns and practices.• A mind set. A Marketing tool.• Web as platform.• User participation.• Content aggregation and
generation.• Hackability and remixability.• Hype – yes.• Substance – yes.
37
Web is a Platform
38
Web as Programming Platform.• Perpetual Beta.• Cooperate, Don't Control.• Design for Mash-ups, hackability,
remixability. Syndicate Content. • Web Services. Standards. Protocols.• Collect the collective intelligence
The Social Web. • Long tail. Network effects by default.
Collective Intelligence.
• Users Add Value. DIY. Tagging. Blogging. Social Networking. Social Information. Social linking. Linking people. Content sharing.
• Unique content by people.• “Perpetual Beta” of syndicated content.
Adapted from Tim O’Reilly’s Web 2.0 2005
40
41
Social Recommendations
Collaborative, Social Tagging
Collaborative, Social Bookmarking
Content Sharing
Application Execution
Mash–up hypertext
• Composites OUTSIDE the application and in the cloud.
• Rich User Experience
• Two way linking in blogs and wikis.
• Matt’s quote
45
• Google AdWords + Google Maps + hostip.info + Microsoft Virtual Earth + NASA + NOAA Weather Service + WeatherBug + Yahoo Geocoding + Yahoo Maps + Yahoo Traffic
47
The Hypermedia NodeWeb 1.0: HTML pages
served up then viewed
using a browser
• Read • Page• Static • Web Coders • Client/Server • Web Browser• Geeks
Web 2.0: Web pages plus other
content, shared (interactively) over the web. More like an application than a page
• Write & Contribute• Post• Dynamic• Everyone• RESTful Web Services• Browser, RSS Reader, App• Mass Amateurisation
From Tim O’Reilly’s Web 2.0 2005
What’s this got to do with Hypermedia?
The Social Writingsharing of content and links.
The Infrastructure for Dynamic Hypertext.Content syndication IS computation in the Hypertext.
Mash-ups ARE composite hypermedia nodes.The Return to Reuse.
Link behaviour in the browser.Two way linking in wikis and blogs.
Link types by microformats. Javascript Libraries for multi-headed links etc.
Web 2.0Microformats AJAX, JSONTagging
CSS, XSLTXPointer, XML
Web 2.0REST, TaggingRSS, ATOM, APIsOpenID, Identity 2.0Ruby on Rails
50
Linking by FolksonomiesWords and Tags
Link Service
Link base
Document LinkedDocument
COHSE: Link Generator Over the Web
COHSE: Linking in the World Wild Web
COHSE: Linking in the World Wild Web
54
55
57
Mind the Gap! Linking is tags, APIs and browser
• Link model?– No information model and information integration
mechanism
• Tags and folksonomies - a bit hit and miss– Controlled vocabularies?– Back in the day we talked about Conceptual
Hypermedia Systems.
58
• Berners-Lee and Hendler vision• Top Down and Planned. • Research agenda for A.I. and
data management• An enrichment of the Web• Web as Knowledge Space.• Open Information Infrastructure• Universal medium for Information
exchange and fusion.• Semantic content aggregation,
integration and generation.• Machine participation.• Hype – yes.• Substance – possibly.
59
Web as Information Platform.• A Layered, Open Information Model • Mark-up and tagging framework.• Uniform Data Model.• Enriching Knowledge Model.• Self-describing, machine processable.• Web Services. Standards. Protocols.• Collect collective intelligence• Infer intelligence
The Socio-Semantic Web. • Network effects by default. Collective
Intelligence. Computed Intelligence.• Users Add Value. Tagging. Blogging. • Machines Add Value. Content sharing.
Content and connection inference.• Unique content by people and automated
processes.• “Perpetual Beta” of semantics.
URI><a href=HTML
Web page Any Web Resource
<a href=“http://…”>
URI
URIURI
Resource Description Framework
The Semantics are in the Link
URI
URIConcept Web Resource
RDFSOWL
Adapted from Figure by Jim Hendler
p -> a; p=a
p -> a; p=a
p -> a; p=a
p -> a; p=a
p -> a; p=a
Web of Trust
Data Fusion
Model FusionR
DF
OW
L, RD
FSSW
RL
XML
SKO
S
Navigate the Semantic Web
Navigate the Web
Topic
Location
Date
University
Company
Funder
Annotation assert facts using terms (metadata in RDF)Represent terms and their relationships (ontology in RDFS/OWL)
ConferenceEvent Organisation
Professional Body
66
Annotation Links + Ontology• Annotation allows explicit decoration of documents
with concepts.
Link Service
Document LinkedDocument
Ontology LinkbaseAnnotations
68
SeaLife Project
69
70
71
Semantic Media Wikis
Link types – Yay! KB within Wiki
Courtesy Max Völkel
73
W3C Semantic Web and HealthCare And Life Sciences SIG
Courtesy Susie Stephens, Lilly
Smashing: A Semantic-Mash-Up
Courtesy Alan Ruttenberg, Science Commons
mc schraefel
The Next Big Thing: Adaptive Web-Based SystemsPaul De Bra, Lora Aroyo and Vadim Chepegin Journal of Digital Information, Volume 5 Issue 1
Article No. 247, 2004-05-27
Web 2.0Microformats AJAX, JSONTagging
CSS, XSLTXPointer, XML
Web 2.0REST, TaggingRSS, ATOM, APIsOpenID, Identity 2.0Ruby on Rails
Semantic Web
81
Web as Platform.• Integration in the model and the browser.• Lightweight development models.• Model, yes. Infrastructure, no.
(http://openanzo.org/).• Content. A little semantics goes a long way (SKOS)
The Socio-Semantic Web. • Consensual semantic acquisition• Social networking, Folksonomies, Tags• Plus ontologies• User Participation models.• More thinking, less reasoning.
Web 3.0?
82
Hypertext?
Web?
Hypertext?
84
“We should make merry and be glad”• Hypermedia as collective intelligence
– Not in the way the visionaries thought.– Really Open. We let in the oiks.
• Open Hypermedia as the mechanism– Composites & Links are thriving. Its done. Get over it.– Why bother with a formal link model? – Versioning and anchor models are still missing….
• Hypermedia authoring by Semantic Model authoring– Was link management, Now RDF management– Was dangling links, Now stale annotation and XPointer
frailty. Was link authoring, now ontology authoring
85
Engineering complexity• The Web worked because it was simple.• Extra layers of software engineering.• Xlink bombed. Need more than a standard.• Web succeeded without a link model.• Keep the footprint small. It will break. Get over it.• Vanilla technologies.
Authoring Complexity. • Extra layers of model. • Content! The Ontology! The Annotation!• Temptation to be too clever.• Hypermedia authoring becomes Annotation authoring –
get your link base!• Hypermedia authoring becomes Ontology authoring.
Stumbles
But ontology authoring and annotation authoring isn’t
hypermedia authoring
Seduced by knowledge …
• The ontology model is seductive way of driving the navigation between resources
• But a Model of knowledge is NOT a Model of Navigation
• You know about Hypermedia, Hypermedia design, Hypermedia patterns, Hypermedia navigation
Hypermedia - This is Our Time Figure Courtesy Tim Berners-Lee
89
Big take homes• Web 2.0 is deeply intertwingled with the hypertext
vision
• Semantic Web isn’t the universal panacea. It has many of the same problems as oh.
• It’s a myth that the Web doesn’t do links.
• Still more to do. You can help.
• Embrace!
90
Hypertext
Web? Web?
91
With thanks to…• Sean Bechhofer, David De Roure• Rest of the COHSE team: Yeliz Yesilada, Bernard
Horan, Les Carr, Wendy Hall• Simon Harper
• http://cohse.cs.manchester.ac.uk/• http://www.biotec.tu-dresden.de/sealife/
92
93
Questions…just what is Open?
• Are first class links worth the bother?– “I don’t want anyone with a GeoCities account
pointing back to me”• Is resource reuse only useful for closed teaching
systems? Or closed enterprises?• Open Hypermedia on the web – open to resources,
closed to the application?• Is the semantic web just link mashing?
94
Prodigal Son Returns• How to help and what to learn from.
• What we did wrong and hence what will go wrong.
• Used mainly in teaching.
• Where do you get the links from?
95
This is why our time has come.
• Back to the future• Your future is my past
• No.
• We are not waiting for the to catch up, the web is waiting for us.
• Could do if it wanted.
• The prodigal son picture, relabelled.