The World Wide Web: Recent developments, The Future and European Industrial Experience

The Great Room, Royal Society of the Arts

Tuesday, 8th December 1998

• Overview: Josef Dietl (W3C)
• XML Update: Jon Bosak (Sun Microsystems)
• XML and Industrial Data: Tony Stewart (RivCom)
• XML in News and Financial Information: Nic Fulton (Reuters)
• RDF: The Resource Description Framework: Janne Saarela (W3C)
• RDF in Digital Libraries: Stu Weibel (OCLC)
• CSS and XSL: Hakon Lie (W3C)
• CSS for WAI: Daniel Dardailler (W3C)
• Vector Graphics on the Web: Chris Lilley (W3C)
• SMIL: Multimedia on the Web: Michael Wilson (RAL)
• TV on the Web: David Bradshaw (BBC)
• The Future: Jean-Francois Abramatic (W3C)

Overview: Josef Dietl (W3C)

The Past

• Joining W3C early on was "an act of faith"
• Last year, the W3C Office in the United Kingdom was opened
• Speakers from W3C only
• Similar events, exhibitions in other European Countries
• W3C has delivered key specifications for the evolution of the World Wide Web
• Speakers from W3C Members talk about their applications of W3C's technologies
• W3C Members are preferred suppliers or expert users

The World Wide Web Consortium

Leading the Web to its full potential

• Since 1995
• Directed by Tim Berners-Lee
• 301 Members
• Four Domains:
  • Architecture
  • Technology & Society
  • User Interface
  • Web Accessibility
• Three Hosts:
  • INRIA (Europa)
  • MIT (The Americas)
  • Keio Univ. (Asien)
• Seven W3C Offices
• ~50 Team (and almost 15 Languages!)

Highlights in World Wide Web Technology in 1998

• XML
• Style Sheets
• Universality: accessibility, access devices, (internationalization)
• The Future

Other topics of the past year

W3C Recommendations

(see http://www.w3.org/TR/)

• Document Object Model Level (DOM) 1
• Synchronized Multimedia Integration Language (SMIL)
• PICS signed label (DSig) 1.0
• Mathematical Markup Language (MathML) 1.0
• ... plus maintenance of seven other W3C Recommendations

W3C Activities

(See http://www.w3.org/)

• Style Sheets
• Document Object Model
• Math
• Graphics
• Internationalization
• Mobile Access
• Amaya
• Digital Signature Initiative
• Metadata
• PICS
• Privacy [P3P]
• Security
• Policy
• Electronic Commerce
• HTTP
• HTTP-NG
• Web Characterization
• Synchronized Multimedia
• XML
• Television and the Web
• Jigsaw
• Libwww
• Web Accessibility Initiative

Acknowledged Submissions to W3C

(see http://www.w3.org/Submission/)

• 20 so far this year
• 16 in 1997
• 3 in 1996

The stage for the speakers

• XML
• Style Sheets
• Universality: accessibility, access devices
• The Future

XML Update: Jon Bosak (Sun Microsystems)

XML and Industrial Data: Tony Stewart (RivCom)

XML in News and Financial Information: Nic Fulton (Reuters)

Any views expressed in this presentation are those of the presenter except where the presenter specifically states them to be the views of Reuters Ltd.

Reuters News Operations

• Reuters is the world's largest news and television agency:
• 2,000 journalists, photographers and camera operators
• 169 bureaux serving 163 countries
• 10,000 stories or 1.5 - 2.0 million words per day
• 25 different languages
• 93 countries use Reuters television news coverage
• 100 million pageviews per month via 140 Internet sites

Reuters Financial Information

• Reuters operates the world's most extensive private communications network:
• 460,000 users in 60,000 organisations
• data from 292 exchanges and markets
• data from 5,000 banks and brokers
• 3,200 updates per second (peak)
• 50 million updates per day

Reuters Technology Aims

• Reuters prides itself on using the latest technology
• Reuters wants to use more off-the-shelf technology
• Our customers want to use standards-based technology

Reuters Technology History

• 1849 - Carrier Pigeons
• 1923 - Pioneers use of radio for news
• 1980s - Satellite delivery
• 1987 - Integrated Data Network
• 1995 - Reuters Web

Reuters and XML - Why?

• XML is the next evolutionary step
• XML can be processed by machine
• XML can be displayed using XSL or CSS
• XML supports Unicode
• XML is self-describing and portable
• XML is supported by everyone
• XML is not a hostage to vendor control

XML and Financial Information

• Financial information is structured data
• XML is great for structured data
• XML is great for financial information

<Record xmlns="http://www.reuters.com" instrument="IBM.N"    currency="USD" type="fullimage">
  <Last wholePart="169" numerator="12" denominator="16"/>
  <Bid wholePart="169" numerator="12" denominator="16"/>
  <Ask wholePart="170" numerator="1" denominator="16"/>  <PERatio value="23.42">
  <Earnings value="24.00">
  <ExDividendDate xmlns:dt="http://www.w3.org" dt:value="1998-11-06"/>
  <Volume value="10634050"/>
  ...
</Record>

XML for structuring

• Namespaces for structure 'ownership'
• XSL or CSS for rendering to a display

XML and News

• News stories are brief
• Users filter or search the stories by topic e.g. foreign exchange, lifestyle
• Stories often have embedded objects e.g. images, tables, graphs
• Information may be associated using links

Using XML for News

• XML for document syntax
• RDF for metadata syntax
• Dublin Core for metadata semantics
• Reuters topic codes for dc:Subject
• 'Embed' a photo with XLink / XPointer
• More stories via XLink / XQL / XPointer

Reuters and XML - Issues

• Update format [XML Fragments]
• Update / change events [DOM events]
• Compression system [HTTP-NG]
• Stylesheet complexities e.g. BID/ASK [XSL]
• XML query language [XQL]
• DTD, RDF Schemas, DCD, XML Schemas
• Inheritance, data typing, enumeration
• Elements vs. attributes vs. values

Summary

• Financial Information is structured data
• News is structured data
• Reuters has LOTS of structured data
• XML is THE off-the-shelf open standard for structured data
• But there are issues.

RDF: The Resource Description Framework: Janne Saarela (W3C)

Outline

• History of RDF
• Inside RDF
• Outside RDF
• XML vs. RDF

History of RDF

• Different applications require different types of metadata
• Platform for Internet Content Selection (PICS)
• Electronic Commerce
• Intellectual Property Rights
• W3C's approach to define an infrastructure for all of them

Inside RDF

The metadata infrastructure defined in two specifications

RDF Model & Syntax specification: http://www.w3.org/TR/WD-rdf-syntax

RDF Schema specification: http://www.w3.org/TR/WD-rdf-schema

RDF Data model

• Resource
• Property
• Value
• subject
• predicate
• object
• Statement

Data model - theory

• multiple representations
• directed acyclic labelled graph
• 3-tuples (triples)
• XML transfer encoding

Data model - 3-tuples

• binary relationships (Universal relation)
• { predicate, subject, object }
• { author, 'http://www.site.net/doc.html', node }
• { name, node, 'Janne Saarela' }
• { email, node, 'jsaarela@w3.org' }

Data model - XML syntax

• Well-formed XML
• Validation through RDF schemas
• No XML/DTD type of validation

<RDF xmlns="http://www.w3.org/TR/WD-rdf-syntax/">
</RDF>
<RDF xmlns="http://www.w3.org/TR/WD-rdf-syntax#">
<Description about="http://www.site.net/doc.html">
</Description>
</RDF>
<RDF xmlns="http://www.w3.org/TR/WD-rdf-syntax#">
<Description about="http://www.site.net/doc.html">
  <a:author xmlns:a="http://www.schema.org/">
  Janne Saarela
  </a.author>
</Description>
</RDF>
<RDF xmlns="http://www.w3.org/TR/WD-rdf-syntax#">
<Description about="http://www.site.net/doc.html">
  <a:author xmlns:a="http://www.schema.org/">
    <Description>
      <a:name>Janne Saarela</a:name>
      <a:email>jsaarela@w3.org</a:email>
    </Description>
  </a:author>
</Description>
</RDF>

Description block

RDF has the requirement of enabling embedding into legacy HTML exposed element content avoided by flexible switch of elements into attributes

<RDF xmlns="http://www.w3.org/TR/WD-rdf-syntax#">
<Description about="http://www.site.net/doc.html"
                      a:author="Janne Saarela"
                      xmlns:a="http://www.schema.org" />
</RDF>

Implications of XML syntax

• One RDF data model has multiple corresponding XML encodings
• One XML encoding has one and only one corresponding RDF data model

Inside RDF - summary

• Data model
  • graph, 3-tuple, XML representation
  • Syntactic sugar of RDF/XML and their implications
  • Higher-order statements
• RDF schemas
  • User-defined types
  • Primitive data types to be defined by XML Schemas WG
  • Source of RDF predicates
  • Constraints on the use of predicates
• Constructing RDF schemas
  • Types are defined as classes
  • a resource may have a specific "type" property
  • Predicates are defined as attributes of these classes
  • a resource of given type may have a given predicate (domain constraint)
  • a resource of given type may be the value of a given predicate (range constraint)

XML vs. RDF

• Could we use plain XML instead of RDF?
• Can XML encode a graph?
• Use ID/IDREFs within one single document
• Can XML support higher-order statements?
• Establish naming and linking convention
• Can XML support embedding in HTML?
• Define interpretation of attributes
• Ah, I just rewrote RDF specification

CSS and XSL: Hakon Lie (W3C)

Why style sheets?

• separate style from content
• quicker downloads
• device-independent document formats
• accessibility

W3C Style Sheets Activity

• 94: Cascading Style Sheets proposed
• 94: W3C Founded
• 96: CSS1 Recommendation
• 97: Extensible Style sheets Language (XSL) proposed
• 98: CSS2 Recommendation
• 98: First XSL Working Draft

Sample CSS

  h3 {
    font-family: Helvetica, sans-serif;
    color: black;
    background: white;
  }

CSS Recommendations

• CSS1
  • simple visual model
  • fonts, colors, layout
  • declarative
• CSS2
  • downloadable fonts
  • aural properties
  • positioning
  • supports XML and HTML
  • media-specific style sheets

CSS Implementations

• 96: Microsoft IE 3
• 97: Netscape N 4, MS IE 4
• 98:
  • Netscape nglayout/GECKO
  • Microsoft IE 5 beta
  • Opera 3.5

Improving CSS implementations

• post-recommendation activities
• CSS1 Test suite
• W3C Core Style Sheets
• CSS Validator
• http://www.w3.org/Style
• www.webstandards.org

XSL

1. tree transformations
   • generates XML (+ CSS)
   • generates HTML (+ CSS)
2. set of XML formatting objects
   • all CSS1/CSS2 properties will be usable
   • focus on print
   • new functionality = CSS3?
   • W3C Formatting Model

XSL Recommendation: expected June 99

News example

Use CSS to:

• select fonts/colors etc.
• turn elements on/off
• simple additions: (London)
• text transformation: (LONDON)

Use XSL to:

• perform transformations
• rearrange elements
• extract headlines
• generate TOCs

W3C Formatting Model

• several specifications expose formatting:
  • CSS
  • HTML
  • MathML
  • SMIL
  • XSL
• one common underlying model
• various languages expose various parts

W3C Style Sheets

• CSS conformance
• XSL adds XML-XML transformations
• common W3C formatting model

CSS for WAI: Daniel Dardailler (W3C)

Context

• WAI is W3C Domain dealing with:
  • Improving W3C technologies (HTML, CSS, XML, DOM, etc)
  • Producing Guidelines for best practice use of Technology
    • Content guidelines
    • User Agent guidelines
    • Authoring Tool guidelines
  • Coordination of Tools development
    • Evaluation
    • Repair
    • Transform
  • Education & Outreach
• WAI helped the design of the CSS2 Recommendation

CSS End-user benefits

• Separation of presentation from content means much easier to adapt the presentation to the users' needs.
• Greatly limit the use of pixel images for textual messages
• Faster to download
• Improve site presentation consistency
• Searchable text
• No more cumbersome author tricks that only some user agents understand.

Some Bad Tricks without Style

• pre, LI or BLOCKQUOTE for indenting and positioning.
• Spaces between letters which confuse non-visual users.
• Spacer "invisible" images used for justification
• Lots of " " to fill out empty table cells
• Complex HTML Table for layout that interfere with logical reading order

CSS2 adds to CSS1

• CSS1:
  • basic syntax, cascading model, classing & formatting
  • font, color, box/margin, border properties
• CSS2:
  • user control of cascading
  • audio rendering
  • positioning
  • downloadable fonts
  • numbering and generated text

Better user control

• CSS1 gives author final say
• CSS2 changes !important to allow end-user to override
• Also introduces inherit to give more flexibility
• Example: force all backgrounds to white and all foreground colors to black:

  BODY { 
  color: black ! important;  
  background-color: white ! important 
  }
  * { 
  color: inherit ! important ; 
  background-color: inherit ! important 
  }
  

Orientation and navigation

• Numbering, Generated text facilitate orientation.
• CSS2 provides a variety of international numbering styles, with flexible style control.
• Example: to number all paragraphs in a document

  p:before { 
  display: marker;                   /* Create a marker  */
  content: counter(parnumber) ". ";  /* for a counter    */
  counter-increment: parnumber       /* and increment it */
  }
  

Aural CSS: ACSS (pronounced Access)

• Support various sound qualities (volume, pause, pitch, etc) and mixing properties
• Equal to font and color variation for non-visual access.
• Example: specify how to speak h3 header

  h3 {
  voice-family: paul;
  stress: 20;
  richness: 90;
  cue-before: url("ping.au")
  }
  

Attribute selectors

• Access to alternative representations of content
• Direct access to alternate text and other attributes
• Example: Element with some attribute set

  IMG[longdesc] { border: solid green }
  

• Example: Access to attribute value in CSS production rule

  IMG:after { content: attr(alt) }
  

Conclusion

• Some items dropped in favor of timeliness
  - Grid unit, Tree transform, Footnote/Endnote
• HTML4 + CSS2 provides overall better Media support
  - braille, embossed, tty are now valid media types
• WAI/CSS Success story of cooperation/coordination

Good Design is Accessible Design

References

• Cascading Style Sheets (CSS), level 2 is a W3C Recommendation available at:
  http://www.w3.org/TR/REC-CSS2.
• The HTML 4.0 Recommendation is available at:
  http://www.w3.org/TR/REC-html40.
• "WAI Resource: CSS 2.0 Accessibility Improvements". Available at:
  http://www.w3.org/WAI/References/CSS2-access.
• Please consult the W3C Web Accessibility Initiative home page at:
  http://www.w3.org/WAI.
• "WAI Page Authoring Guidelines" for designing accessible documents at:
  http://www.w3.org/TR/WD-WAI-PAGEAUTH.

Vector Graphics on the Web: Chris Lilley (W3C)

No set of slides provided

SMIL: Multimedia on the Web: Michael Wilson (RAL)

Talk Outline

• Example SMIL Presentation
• The SMIL Language
• SMIL Tools
• SMIL take-up

Available Technologies

• HTML/XML - no timing
• DHTML (Java, CSS, DOM) - considerable programming effort, DOM standard
• CD authoring tools - plug-in only
• convert everything to video stream - RealVideo/MS NetShow
• New W3C multimedia language - SMIL

SMIL Objectives

• Support 5 Constructs
  • layout
  • content
  • control/links
  • timing/synchronisation
  • tailoring
• Became W3C recommendation June 1998

SMIL Language

• Defined as an XML DTD
• declarative language like HTML, easier to learn and use than (DHTML: Java and DOM and CSS)

<smil> <head> ... </body></smil>
<layout>
<region id="V-Main" left="3%" top="44%" width="46%" height="40%"/>
</layout>

Media Content - http, rtsp, etc., local file

<video region="R1" src="video.avi"/>
<img src="http://image.jpg"/>, 
<audio src="rtsp://sound.wav"/>, 
<text src="text.html"/>, 
<textstream /> e.g. tickers
<animation />
<ref /> - other media types

Media Timing/Synchronisation

<par> 					 
	<seq>
			<img src="slide1.jpg" dur="10s" />
			<img src="slide2.jpg" />
	</seq>
			<audio src="english.wav"/>
</par> 					

40+ attributes to set duration, synchronisation etc..

Control

WWW, therefore limited to hyperlink navigation

<a show="new" href="./file1.smil#1">
		<video id="Anchor" region="V-Main" src="/mpeg/file2.avi"/>
</a>

• also <anchor> statement for video anchors that are limited in both time and space
• jump into or out of continuous media anchors
• no anchor for audio - use scrolling text captions

Tailorability

<switch> statement and test attributes
<switch>
 <audio src="french.wav"  system-language="fr"/> 
 <audio src="english.wav" system-language="en"/>
</switch>

• select media content for different user languages , screen size, screen colour depth, comms bitrate (e.g. GSM vs LAN)
• optional captions for users with hearing difficulties
• optional dubbing for users with visual difficulties
• different layouts for screen size (e.g. GSM phone).

SMIL Tools

• Browsers/Players
• Editors/ Authoring Tools

SMIL Browsers

• HPAS, Netscape Plug-in, Digital research tool only
• Helio, Java Applet player - very limited
• Productivity Works, LP Player Audio Only, DAISY compatible
• Cresendo Forte - audio only- music
• Real Networks G2 Player
• Chameleon GRiNS Player

Authoring Tools

• Tags - Allaire SMIL Integration Pack
• Audio only - LP Studio
• Template filling - RealMedia SMIL Presentation Wizard
• SMIL compatible multimedia editors
  • G2 Objects for Macromedia Dreamweaver
  • Digital Renaissance T.A.G. Editor
  • Sausage Software, SMIL SuperTool
  • Veon V-Active
  • GRiNS Authoring Environment

The Veon V-Active Editor

• The tool includes templates designed to facilitate the production of SMIL presentations.
• V-Active will not only encode audio and video, it features an integrated timeline utility.
• Produces <anchor> statements for video objects.

GRiNS Editor

• 5 views for 5 constructs
• Layout
• Hierarchy
• Timeline
• End User
• Links

SMIL Advantages

• Not a full programming language - DHTML
• Declarative language - easy to learn
• Easy to convert to/from other languages - e.g., HyTime
• Controls whole player unlike CD-Rom tool plug-ins
• Less bandwidth than digital video don't stream whole image, only continuous parts
• Commercial tools already exist
• Over 50 major media users use SMIL including CNN, Fox OnLine News etc.
• W3C Recommendation, not proprietary

Risks to SMIL Positioning

• Are SMIL layout and linking sufficiently consistent with XML, CSS and HTML?
• Should alternatives be addressed at a more general level than SMIL?
• Is a declarative language enough, or is scripting required by experts?
• Microsoft may promote HTML+TIME instead of SMIL, but RealPlayer G2 will ship with AOL V4.0

SMIL - The Future

• SMIL SIG exists now to produce errata
• SMIL WG is proposed from February 1999 to 2000
• Extensions of SMIL 1.0 e.g. transition effects, external clock synchronisation, streamed events, anchor indication by cursor, integrating Xlink
• Modularisation of SMIL 1.0
• DOM for documents with SYMM
• Integration of SYMM functionality and HTML

Conclusions

• SMIL became W3C recommendation for synchronised media on WWW in June 1998
• XML technology is a key component of SMIL
• More efficient than alternative formats
• SMIL is easy to learn and use
• There are risks to the positioning of SMIL
• Example browsers and editors available now

TV on the Web: David Bradshaw (BBC)

The Future: Jean-Francois Abramatic (W3C)

• Why: Goals
• Who: Actors & Roles
• How: Process
• What: Achievements
• Next: Challenges

"Leading the Web to its full potential"

Provide the Infrastructure for the Global Information Society

• Prevent Market fragmentation
• Develop a powerful Architecture
• Guarantee Accessibility
• Develop best Practices

using Collective Design

W3C: Actors & Roles

• Industry-driven: Members
  • Software & Systems suppliers
  • Communication suppliers
  • Content providers
  • Large end users
  • Research laboratories
• Architecture-driven: Team
  • Roadmap by Tim Berners-Lee, Director
  • Identify & Manage Dependencies between Activities
  • Manage the Process to reach consensus
  • Promote the results
• Society-driven: Public
  • General feedback on Societal issues
    • Accessibility
    • Trust
  • Continuous feedback on Deployment issues
  • Early feedback on Working Drafts

W3C: Hosts & Offices

• Hosts
  • MIT: October 1994
  • INRIA: April 1995
  • Keio: August 1996
• Provide
  • Management Team
  • Engineering Team
  • Communication Team
    • Web site
    • Events
    • Newsletter, Newswire
    • Press relations
• Offices
  • Europe: UK, Germany, Netherlands, Sweden, Greece, Italy (1Q99)
  • Asia: Hong Kong, Taiwan
  • Africa: Morocco, Tunisia (1Q 99)
• Provide
  • Communication
    • New membership
    • Members contact
  • Mirroring & Translating Web site
  • Feedback from local needs

W3C: Process

• Ideas & Issues
• Submissions
• Activity proposal
  • Scope
  • Dependencies
  • Organization
• Working Groups & Interest Groups
  • Mailing list
  • Web area
  • Conference calls
  • Face-to-face meetings
• Documents
  • Notes
  • Submissions
  • Working Drafts
    • Member-only
    • Public
    • Last Call
  • Recommendations

W3C: Achievements

• HTML 3.2, HTML 4.0
• HTTP 1.1
• CSS1&2
• XML 1.0
• SMIL 1.0, MathML 1.0
• PNG
• DOM Level 1
• PICS
• ....

W3C: from Achievements to Deployment

• Sample code
  • Client: Amaya
  • Server: Jigsaw
  • Library: Libwww
• Validation tools
  • HTML
  • CSS
• Development Forum
  • HTTP 1.1
• Education & Outreach
  • Accessibility
  • Privacy

W3C: Work in Progress

• Metadata (RDF)
• Digital Signatures (DSig)
• Privacy (P3P)
• Intellectual Property Rights
• Structure (XML, HTML)
• Graphics & Multimedia (SVG, SMIL)
• Style (CSS, XSL)
• Accessibility (WAI)
• New Generation Protocols (HTTP-NG)
• ... from Member input

Convergence

• Television and the Web
  • Workshop in Sophia-Antipolis (June1998)
  • Interest Group (September1998)
  • Web components: HTML, CSS, Addressing
• Mobility and the Web
  • Workshop in Tokyo (April 1998)
  • Interest Group (June 1998)
  • Agreement with WAP Forum (Wireless Application Protocol)
  • Web components: HTTP, HTML, CSS

Thank you

More information at http://www.w3.org/

Feel free to join!