when?
Post on 17-Oct-2014
4.197 views
DESCRIPTION
TRANSCRIPT
When?
Dan Brickley <http://danbri.org/>
‘When...?’Semantic Web technologies: where next...
Basket Cases
• Lessons (things that happened already)
• Trends (things happening anyway)
• Requests (things to make the future sooner)
Lesson
Trend
Request
• the Semantic Web project: history, hypertext & RDF
• Two ‘claims & hypertext’ scenarios
• Awkward Bit: learning our lessons
• play to our strengths: aggregation
• Trends in UI, search and services
• Big picture: Web as the new public record
Overview
Lesson Trend Request
HISTORY
HYPERTEXT
& RDF
1.
Geoffrey Sneddon, Tim Berners-Lee, William ‘When?’ Loughborough.
W3C Technical Plenary meeting, 2008, Cannes, France.
Douglas Engelbart’s 1968 Demo
1968
1989
“To a computer, the Web is a flat, boring world, devoid of meaning. This is a pity, as in fact documents on the Web describe real objects and imaginary concepts, and give particular relationships between them.”
“For example, a document might describe a person. The title document to a house describes a house and also the ownership relation with a person.”
“Adding semantics to the Web involves two things: allowing documents which have information in machine-readable forms, and allowing links to be created with relationship values. [this will] help us exploit the information to a greater extent than our own reading.”
Tim Berners-Lee "W3 future directions" keynote
1st World Wide Web Conference Geneva, May 1994
1994
!
!
2008
The World:
The Web:
Pages make claims
ongoing.org: this is the OpenID for Tim. Tim knows Henry.
ongoing.org: Tim’s workplaceHomepage is http://sun.com/
bblfish.net: Henry workplaceHomepage is http://sun.com/
bblfish.net: Henry knows Tim. Henry’s phone number is, [...]
Trend
Strengths & Weaknesses
• Rich mixing, within and between documents
• any RDF data can use any RDF/OWL vocab
• can mention and describe anything related
• can omit information without breaking
• Things can be identified in multiple ways
A lot to be proud of...
• world-class, Web-scale data mixing environment
• current best bet for treating the Web as data
• optimised for massive decentralization
• somewhat chaotic, semi-structured, scruffy
• but grounded in formality and simplicity
CLAIMS
2.
&HYPERTEXT
1. Express the basic claim in RDFa.
http://www.w3.org/People/Alumni
2. Hyperlinking for discovery from W3C homepage.
eg: <a href=”/People/Alumni” rel=”xyz:alumniPage”>alumni</a>
Trust comes from linked information: claims in context.
“Does the Web know what you’ve been doing?”
Request
Problem partition
• what is W3C’s homepage?
• we can check with Wikipedia/DBpedia
• what data/documents do they provide?
• via RDFa, RDF/XML, GRDDL, SPARQL
• what do they tell us? /People/Alumni etc.
• are they up to date? reliable? risky? wikis?
• a case-by-case decision
foaf:tipjar
Revisiting the “oh yeah?” problem with OpenID authentication.
Trend
Trend
BBC Music: view source
Q: if we know their homepage and myspace page...
...how does that change what you’ll entrust to these OpenIDs if they log into your site?Request
Problem partition
• what is the Rumblestrips homepage?
• what is their myspace page?
• is the OpenID I’ve just seen, one of theirs?
• so that tipjar page linked from their FOAF, I can trust it?
• ‘dunno mate’.
Request
THE
AWKWARD
BIT
3.
Things that might’ve worked out better.Lesson
The art of compromise...
RDF: data spork?
Why would Mozilla walk away?
LessonLessonLesson
Redland+SQLLite+Mozilla = ?
Lesson
Lesson
Lesson
(clue: wanted - stable, packaged pure-Ruby RDF toolkit)
Why would a startup with an RDF-guru CTO not do RDF import?
Lesson
Learning Lessons
Read/write is harder than read-only aggregation.
Tools. Packaging. Testing. QA. Documentation.
Redland (RDF in C) is important.
So are the scripting languages. All of them.
Lesson
Request
Request
Request
Lesson
Next time around...
• Drupal 7
• OpenOffice.org
• KDE
• XMP
• Yahoo! SearchMonkey, Google SGAPI, ....
• (not to mention widgets, oauth etc.)
Request
Playing to our strengths?
Primarily, I think this means aggregation not management of data. This follows from the necessarily
patchwork, open-world nature of our data model. And of the Web.
RDF is beautiful for aggregation; challenging for data management.
discuss...
Request
GRDDL, SQL-to-SPARQL, RDFa, ...
Strengths as weaknesses
• Rich mixing, within and between documents
• any RDF data can use any RDF/OWL vocab
• can mention and describe anything related
• can omit information without breaking
• Things can be identified in multiple ways
...being greater than the sum of your parts?
The strange appeal of the Semantic Web...
TRENDS
IN
UI, SEARCH & SERVICES
4.
Trend
What will ‘it’ be like?
An invisible Semantic Web...Somewhat better search results.
Relevant people grouped in addressbook.
A calendar full of useful information.Long-term archival tools.
Slightly smarter spam filters.
Auto-maintained cross-references.
A quiet revolution?Trend
What, no 3D flythrough?
Freebase Parallax
• From the skyscrapers in Hong Kong
• find the architects
• take the other buildings they made
• plot those on a map
Trend
(from David Huynh, ex-SIMILE)
<http://code.google.com/p/freebase-parallax/>
• or take US presidents
• the ones who are republican
• their sons and daughters
• the schools they went to
• the inevitable map...
• or timeline, spreadsheet, blog post...
Trend
Sets and links
• The projects you work on
• The people who work on them too
• their latest public bookmarks
Trend
Two flavours
• skeptical: caring who claimed what
• trusting: navigating flattened information
• dataset selection & the “Oh yeah?” button
Request
Trend
Request
Related UI we’re more familiar with: Google earth layers, overlaying calendars, Photoshop layers...
More set-based navigation: gfacet
“which cities are the bands that wrote this music from?”
Trend
Trend
Data :)
Search & Services
Yandex.ru, Aug 15th 2008:
“The FOAF (friend of a friend) standard makes blog search or social network search deeper and more accurate, in particular, it allows searching friend feeds and user profiles”
“the largest blog services in the Russian internet including Livejournal.com, Liveinternet.ru and Blogs.Mail.ru, represent user profiles in FOAF.”
Trend & Sindice, Garlik Qdos, Falcons, ...
Is Google the next Google?
“The Social Graph API makes information about the public connections between people on the web more easily available.”
“...indexes the public Web for XHTML Friends Network (XFN), Friend of a Friend (FOAF) markup and other publicly declared connections. By supporting open Web standards for describing connections between people, web sites can add to the social
infrastructure of the web.”
Trend
SGAPI: Microformats & FOAF/RDF
Claim graph analytics:epeus.blogspot.com: “kevinmarks.com is me!”
kevinmarks.com: “epeus.blogspot.com is me!”
danbri.org: Kevin Marks’ site is http://kevinmarks.com/
danbri.org: his foaf:workplaceHomepage is http://google.com/
Request
5.
THENEW
PUBLICRECORD
Who, what, where and when, in...
• Healthcare
• Libraries
• Arts
• History
• Museums
• Education
• Archives...
• Government
• Journalism
• Science
• Law
• Trade
• Research
• Agriculture
an endless list, infinitely interconnected...
Already happening
• The Web is becoming our common public record
• the Semantic Web’s mission is to defragment it
Aside: privacy by obscurity is going away fast. Facebook & ‘social graph’ tools are educating a generation about “public”...
Libraries revisited: FRBR
Request
Trend
Functional Requirements for Bibliographic Records
FRBR:
Credit: William Denton <://www.miskatonic.org/library/2008ola/>Request
Trend
Libraries revisited, ...☑Catalogues.☑Name authority files.☑Thesauri.☐ Classification schemes.
Request
Trend
Keep an eye on: http://openlibrary.org/
Douglas Engelbart’s 1968 Demo
Trend
Re-use
Re-cycle
Anything but seamless.
Reclaim
Repair
Recap
LessonWhen you’re a data mixing system, be a data mixing system.
Lesson
There is plenty more code to write, test and document.
Lesson
Learning from non-‘true believer’ users will speed up adoption.
Stating the obvious?
Trend
Web increasingly machine-readable -XML, µfJSON, MFs, SQL, & -yes- RDF/OWL.
Pages make claims.
RDF UI ideas are maturing.
Massive datasets are being linked.
Integration here plays to our strengths...
Request
1. Get the missing code written. Tested. Packaged. Please!
2. Think about the Web as a linked information system.
3. ...and the Semantic Web as a project not a thing.
Questions?