Brian Kelly UK Web Focus UKOLN University of Bath BA2 7AY B.Kelly@ukoln.ac.uk

1

WWW 6 Trip ReportReport on the Sixth International WWW Conference held in Santa

Clara on April 7-11, 1997

Brian Kelly

UK Web Focus

UKOLN

University of Bath

BA2 7AYB.Kelly@ukoln.ac.uk

http://www.ukoln.ac.uk/

Contents• About WWW 6• W3C Session - HTML• Microsoft IE 4.0• W3C Session - HTTP• Web Site Mapping Workshop • XML Tutorial• Technical Papers• Plenary and Panel Sessions

2

UK Web Focus

UK Web Focus:• National Web coordination post• Responsibilities include:

– Technology watch– Information dissemination– Coordination– Representing JISC on W3C

• Based at UKOLN, University of Bath• Brian Kelly appointed on 1st Nov 1996• Formerly worked at Netskills (Newcastle University) and

universities of Leeds, Liverpool and Loughborough

3

About WWW 6WWW 6:

• Held at the Santa Clara Convention Centre• Workshops and tutorial sessions on Monday

7th April• Technical papers from 8-10th April• Developer's Day and History Day on 11th April• Various other BOFs and meeting (including

Web Accessibility Initiative day)• About 1,800 participants (down on previous

years)

4

Before the Conference• Conference details

including online booking available on Web - http://www6conf.org/

• Online conference (ICE) available for delegates before, during and after conference

• Conference proceedings available online http://proceedings.www6conf.org/

http://ice.www6conf.org/

5

Accessibility

http://www.lbc.co.uk/

Accessibility was conference theme

Web Accessibility Initiative:• About 80 participants• WAI approval

Issues• Accessibility is important• WAI will address:

protocols, software developers, information providers and end users

• Need for university / departmental / project policy?

6

Accessibility

Comments made by Peter Bosher (RNIB):• Poorly designed navigation (when you follow

several links, and then get back to where you started) is a particular problem for the blind who don't have the visual clues that the sighted have

• When navigating the Hotwired site his browser said "link link link link image image image without caption"

• Tables are difficult to process by speech browsers

• Painstakingly competing a form and then getting an error message because the form was browser specific is annoying

7

Further Information

Further information on the Web Accessibility Initiative, and on general accessibility resources is available at:http://www.w3.org/pub/WWW/Disabilities/

http://weber.u.washington.edu/~doit/Other/design.html

http://www.igs.net/~starling/acces.htm

http://trace.wisc.edu/

http://access.adobe.com/

http://cs.cornell.edu/home/raman/

8

HTML Developments

Cougar:• Code name for next version of HTML

• Needed as "forces driving evolution of HTML are threatening media independence"

• First draft available summer 97

• Some parts already released:

WD-htmllink WD-script

WD-frames WD-forms

WD-object WD-entities

WD-fileupload (WD-printing, WD-positioning)

• See http://www.w3.org/pub/WWW/MarkUp/Cougar/ Day 1 - 2 pm

W3C Session

9

WD-htmllink

WD-htmllink• Hypertext links in HTML. • Defines link relationships (e.g. <LINK REL=Next>

could be used to facilitate printing in a single document split into several files)

• Related to Style Sheet work

<LINK REL=Stylesheet MEDIA=print HREF=ukoln-print.css>

• Define TITLE attribute for hints for accessibility e.g. <A HREF="welcome.html" TITLE="The Welcome page for ACME Ltd">text</A>In this example a browser could display the text in, say, a pop-up window.

Day 1 - 2 pm

W3C Session

10

WD-htmllink

WD-htmllink (continued)• Defines uses of META tag (e.g. for use with

PICS)• Define language(s) for documents • Define pages which can be indexed by robots

<META NAME="ROBOTS" CONTENT="NOINDEX, NOFOLLOW">

prohibits robots from indexing and following links

Day 1 - 2 pm

W3C Session

11

WD-script

WD-script• Client-side Scripting and HTML• Defines mechanism for embedding scripting

languages in HTML:<SCRIPT type="text/javascript">document.write ("<EM>This will work</EM>")</SCRIPT>

•<SCRIPT SRC="url"> used for external scripts• Default script language can be defined:

<META HTTP-EQUIV="Content-Script-Type" CONTENT="text/tcl">

Day 1 - 2 pm

W3C Session

12

WD-frames

WD-frames• Defines existing usage of FRAME tag• Includes IFRAME proposal for inclusion of frames

inline in body of HTML documents (FRAME tag replaces BODY tag):<BODY><P>This document contains an inline frame <IFRAME src="foo.html" width=400 height=500> that unfortunately your user agent doesn't support. Alternatively you can get the related document <A href="foo.html">here </A>.</IFRAME> That's all folks!</BODY>

• Initial release of FRAMES caused problems,and there are still accessibility concerns

Day 1 - 2 pm

W3C Session

13

CSS

Day 1 - 2.45pm

W3C Session

The CSS1 (Cascading Style Sheet) spec:• Replaces all HTML extensions• Can replace most uses of tables• Provides floating text elements• Provides control over background

WD-style• Defines relationships between HTML document

and stylesheet• Cascading style sheets can be defined by:<LINK REL=stylesheet HREF="corporate.css"><LINK REL=stylesheet HREF="techreport.css">

• Can define alternate style sheets for media types:<LINK REL=stylesheet MEDIA=screen …>

14

CSS - The Next Wave

In next version of CSS:• Multiple media (using @media)• Improved printing support (headers, footers,

margins, etc.) See WD-printing• Better control over positioning (e.g. out-of-flow

elements, navigational bars, layering) See WD-positioning

• Better font control (mixed fonts - e.g. helvetica and cyrillic for, say, language dictionaries)

• Aural cascading style sheets. See WD-acss For visually impaired an incar use, industrial and medical systems, entertainment, illiterate:

– <STRONG> rendered as loud voice, pause-before– Left column in table spoken in left speaker Day 1 - 2.45 pm

W3C Session

15

CSS Questions

Q How closely matched are the CSS and XML?

A They are a good match

Q Features such as side bars (e.g. <P CLASS=side>) are likely to be widely used. Should such attributes be registered?

A It's likely that a set of standard style sheet definitions will be released, and they'll be widely used

Q Will CSS held to reduce network bandwidth?

A Yes, see http://www.w3.org/pub/WWW/Protocols/NL-PerfNote.html

Q What about headers and footers?

A In next release

Day 1 - 2.45pm

W3C Session

16

HTML Math

The HTML Math spec:• Released in May 97 - see http://www.w3.org/TR/WD-math

• Core standard covering presentation and markup

• Will contain 20 presentation tags with 40 attributes

• Will contain 50 content tags (roughly equivalent to functions on scientific calculator)

• Initially implemented using embedded elements (ActiveX, Java) enabling 3rd parties to develop rendering tools (won't need to wait for Netscape / Microsoft)

• Software will be available in June 97

• Full implementation requires better browser APIs (e.g. DOM)

Day 1 - 2 pm

W3C Session

17

Dynamic HTML

• Based on work of the Document Object Model (DOM) working group

• Provides an API (applications programming interface) for HTML page structure and style

• Microsoft's Internet Explorer 4.0 implements many features of Dynamic HTML

• See:–http://www.microsoft.com/gallery/files/html/default.htm

–http://www.microsoft.com/workshop/prog/aplatfrm/dynhtml-f.htm

Day 1 - 2 pm

W3C Session

18

Microsoft IE 4.0

Presentation in the Industrial Presentations track on Internet Explorer's implementation of Dynamic HTML

HTML is limited:• Animation using animated GIFs is limited• Can't position or layer elements• The web is slow - interaction with server needed

Dynamic HTML:• Based on emerging W3C proposals• Covers:

– Dynamic styles and dynamic contents– CSS positioning– Data binding

Day 2 - 4 pm

19

Microsoft IE 4.0Dynamic HTML

• Positioning control (in X, Y and Z planes)

• Position can be changed at run time (e.g. on mouseOver event)

Day 2 - 4 pm

Dynamic ContentThe Document Object Model (DOM) can be used to change the content at runtime:function change() "new header"...<H1 id=foo onclick=change()>Old header</H1>

For example a table of contents could be built dynamically

20

Microsoft IE 4.0Dynamic Style Sheets

CSS properties can be changed at run time:

<H1 onMouseOver {colour: red;} onMouseOut {color=yellow}>

A more elegant way is to store the Javascript code in the document HEAD (or externally)

CSS PositioningElements can be positioned absolutely or relative to each other

<IMG SRC="smiley.gif" : top: 25% ; zindex=1>

<DIV zindex=2>Text on top of image</DIV>

Day 2 - 4 pmText on top of image

21

Microsoft IE 4.0

Data BindingCurrently data binding (merged data from database with HTML code) is done on the server

Data binding proposal enables it to be done on the client so that, for example, the output from a search engine can be resorted on the client)

But will this be in Cougar? Day 2 - 4 pm

http://www.microsoft.com/gallery/files/html/repeat.htm

22

Microsoft IE 4.0

Dynamic HTML Multimedia EffectsUses IE4 control to provide multimedia effects

Removes need for animated GIF

Variety of effects available:• Transitions, filters,

structured graphics, sequencing

• sprite control (scripted animated GIFs)

• Example - e rotating in 3DDay 2 - 4 pm

How will Powerpoint be marketed if this is possible in HTML?

23

Microsoft IE 4.0

Document Object Model (DOM)• DOM enables every HTML element to be

processed as an object (i.e. DOM is an API for accessing HTML elements)

Day 2 - 4 pm

• DOM is not a set of tags or a new language (it is language independent)

24

Netscape

Netscape have also expressed support for Cougar

See• http://www.netscape.com/focus3/comprod/columns/intranet/open_standards.html

• http://www.netscape.com/flash1/comprod/products/communicator/

25

Cougar and Browser IssuesStandards

• IE 4.0 conforms to HTML 3.2 and CSS1, and implements new W3C drafts (e.g. DOM, forms)

• Netscape have also expressed commitment to Cougar

• Javascript is being standardised by ECMA

Support for legacy browsers?• Don't use new features• New features degrade gracefully• New features won't work (e.g. Tetris example -

10K HTML file, which used no ActiveX controls)

Day 2 - 4 pm

26

Support for Multiple Browsers

How do we deploy new features when there are many different versions and many browsers to support?

• Offer choice at client:Click here if you have IE 4 or Netscape 4 otherwise click here

Maintenance of resources can be done manually or using site management tool

• Offer choice at server:if useragent=IE4.0 then serve index.dom.html else serve index.html

• Various toolkits can provide this (Microsoft Active Server Pages, Netscape Suitespot, PHP/FI)

27

Developer's Day - HTTPTransparent Content Negotiation (TCN)

• IETF draft• Enables:

– Deployment of new data formats and tags– Tailoring of content for new platforms– Internationalisation

• See ftp://ftp.ietf.org/internet-drafts/draft-ietf-http-negotiation-01.txt

• TCN spec server sends list of variants:

Day 4 - 9 am

<IMG SRC="logo"> {logo.gif type image.png}{logo.png type image.png}HTML source

Variant list sent by server

28

Developer's Day - HTTP

Feature Negotiation• IETF draft• Part of TCN, addressing extensibility• No more "Click here from frames, here for

tables" buttons• Universal agreement on new features not

needed• See ftp://ftp.ietf.org/internet-drafts/draft-ietf-http-feature-reg-00.txt

Day 4 - 9 am

29

Web Site Mapping Workshop

Half day workshops on web collections:• Defining relationships between groups of related

documents

• Useful for:

– printing

– off-line browsing

– indexing

• Presentations on:

– Protocols (Web Collections, MCF, Dublin Core)– Applications (HotSauce, WebCutter,

WebMapper)

Day 0 - 9 am

30

Web Visualisation

Various proprietary ways of visualising web sites are available (e.g. MAPA, WebCutter)

Need a standard for defining relationships between web pages, to provide application independence

http://dynamicdiagrams.com/

http://www6.nttlabs.com/HyperNews/get/PAPER40.html Day 0 - 9 am

31

ConclusionsThe Web Site Mapping workshop agreed:

• Split protocol into three:

1 Metadata Model 2 File Format Syntax

3 Metadata Manipulation Language• Further discussions needed on:

– Is a new query language is needed (e.g. SQL, HyTime Query Language)

– Do we need a metadata query language or a general document query language

– A list of common site mapping operations

• XML is likely to be the preferred syntax• Web Collections spec to be updated - http://www.w3.org/pub/WWW/TR/NOTE-XMLsubmit.html

Day 0 - 9 am

32

Metadata Architecture

Much work in progress in developing a metadata architecture for the web

Metadata

Web Collections,

PICS, TCN,

MCF, DSig,

DC, ...

AddressingURL

Data formatHTML

TransportHTTP

33

XML Tutorial

XML:• Extensible markup language• An SGML-lite designed for the Internet• Developed by the SGML community• Tools being developed by SGML tool vendors• Microsoft involvement • Two XML draft specs:

– XML– XML-link A richer form of linking

• Stricter than HTML (to reduce client processing):– Attribute quoting <IMG SRC="logo.gif">– End tags needed <P>A paragraph</P>

Day 0 - 2 pm

English

French

34

What's It Look Like?

What does an XML document look like?

<?XML VERSION="1.0" RMD="ALL" ENCODING="UTF-8"?><!doctype titlepage system "typo.dtd" [<!entity % active.links "INCLUDE">]><titlepage><whitespace type="vertical" amount="36"/><title font="Baskerville" size="24/30"

alignment="centered">Hello, world!</title><whitespace type="vertical" amount="12"/><!--* In some copies the following decoration is hand-colored, presumably by the author *--><image location="http://www.foo.bar/fleuron.eps"

type="URL" alignment="centered"/><whitespace type="vertical" amount="24"/><author font="Baskerville" size="18/22" style="italic">

Munde Salutem</author></titlepage>

<?XML VERSION="1.0" RMD="NONE"?><conversation><greeting>Hello, world!</greeting><response>Stop the planet, I want to get off!</response></conversation>

Simple XML document

More complex XML document with DTD specified

35

The Jumbo XML Browser

Jumbo is an XML browser written in Java.

It was developed to view CML (Chemical Markup Language) resources.

It can view other XML applications. http://www.venus.co.uk/

omf/cml/download.html

36

CDF

XML can be used to define:• Document structure• Structure for metadata• Structure for applications

CDF (Channel Definition Format) is:• An XML application• A proposed standard for push

technology• Developed by Microsoft and

implemented in IE4• See http://www.w3.org/TR/NOTE-CDFsubmit.html

37

Further Information on XML

For further information see: http://www.textuality.com/xml/

http://www.w3.org/MarkUp/SGML/Activity

http://sunsite.unc.edu/pub/sun-info/standards/xml/why/xmlapps.htm

http://www.ucc.ie/xml

http://www.webreview.com/97/05/16/feature/

38

Extending HTML in a Principled Way with DispletsThis paper describes a Java solution to adding new tags (charts, links, maths, etc.) to HTML.

<TAG NAME=CHART .. SRC=chart.class>…</TAG>

<CHART TYPE=BAR>

<TABLE>

<TH> Smith<TR><TH>125…

</TABLE>

</CHART>

Jan Apr Jul OctSmith 125 257 327 250Green 137 140 110 160

HTML source HTML output

Jan Apr Jul Oct

Day 1 - 11am

Technical Paper

39

Extending HTML in a Principled Way with DispletsOther extensions:

Multi-way links

Graphs

Issues:How does this fit in with XML?

How does this fit in with the HTML Math work, which proposes using Java to display Maths output?

Further Information:http://proceedings.www6conf.org/HyperNews/get/PAPER155.html

http://pegasus.rutgers.edu/~chchiu/displets.html

40

Towards a Multimedia World-Wide Web Information Retrieval Engine

Paper that integrates searching for text with searching for images.• Search for "baseball

player"• Select one of the images

retrieved, and search for others similar to this one

• Give schematic drawing (as shown)

• Provide 3D visualisation of search results using VRML

Find an image that looks like this will find Mickey Mouse from the Disney Web site

Day 1 - 4 pm

41

Towards a Multimedia World-Wide Web Information Retrieval Engine

The software is called AMORE.

Prototype available at

http://www.ccrl.neclab.com/amore/

See http://www6.nttlabs.com/HyperNews/get/PAPER3.html

Day 1 - 4 pm

42

Dynamic Reference Sifting: A Case Study Described Ahoy! - a Web application used for finding personal home pages based on recognition of likely hits from directory naming conventions.

Day 1 - 4 pm

See http://www6.nttlabs.com/HyperNews/get/PAPER39.html and http://www.cs.washington.edu/research/ahoy

43

Parasite: Mining Structural Information on the WebBased on heuristic assumptions including:

Hypertext Linking• A linked page is likely to be on the same topic as the

original page (esp. for Yahoo type resources)

Directory Structure• A URL containing a directory below a personal home page

(PHP) is likely to be authored by the person identified in the PHP

Page Structure• Links "near" each other on a page are likely to have similar

topics

Day 1 - 4 pm

44

Parasite: Mining Structural Information on the WebUsed these assumptions to propose applications for finding:

• moved pages• related pages• people

Comments• Brute strength approach• This and preceding paper (on Ahoy!) show

importance of directory naming conventions (directory names provide metadata - what can we guess from the URL www.cs.acme.edu/staff/jsmith)

Day 1 - 4 pm

45

WebCutter: A System for Dynamic and Tailorable Site Mapping

Paper on IBM/Lotus product for website visualisation

• Implemented in Java• Map generated on

the fly • Map can be edited

by end user (to define user's view of web site)

Day 3 - 2 pm

http://www6.nttlabs.com/HyperNews/get/PAPER40.html

46

WebQuery

Paper on various visualisation techniques for searching.

Based on structural information to find "hot spots"

http://www6.nttlabs.com/HyperNews/get/PAPER96.html

Day 3 - 2 pm

47

Transforming Command-Line Driven Systems to Web Applications

This paper described work which used Java to provide access to a legacy command line application (a monolithic Fortran 77 program) on the Web:

• Input to application via Java form

• Output in Java applet window

• Commercial product available soon

Day 3 - 2 pm

http://www6.nttlabs.com/HyperNews/get/PAPER41.html

48

Responsive Interaction for a Large Web Application The Meteor Shower Architecture in the WebWriter II Editor

An HTML authoring tool with server and client side processing.

Makes use of frames.

Uses <BLINK> to define flashing cursor

Day 3 - 2 pm

http://www6.nttlabs.com/HyperNews/get/PAPER86.html

49

Seamless Integration of Interactive Forms into the WebThis paper described limitations of existing forms on the Web and described how Dynamic Forms would overcome these limitations.

Dynamic Forms is based on Java.

But what about W3C work on HTML (new DOM and Forms specs)? http://www6.nttlabs.com/

HyperNews/get/PAPER83.html

50

Plenary Talk

All Authored Works Online: A Global Infrastructure for Universal Access to Information

Raj Reddy, Dean of the School of Computer Science at Carnegie Mellon spoke on the Universal Library project - when "All Authored Works of the Human Race will be available to anyone in the world instantaneously".

His talk is available at http://www.ul.cs.cmu.edu/rr/aaw/aaw.html

51

Web Design Panel Session

Good Web Design Panel Session• A very popular session on Web design• Panelists included Jakob Neilson (Sun),

David Seigel (a leading web designer and author of the 1 bit clear GIF used to provide spacing), Cathy Gill (HTML Writer's Guild) and others

• Strong disagreements between the design community (David Seigel) and the structuralist community (Jakon Neilson)

52

What is Good Web Design?What is Good Design?

• No frames - Jakob Neilson's response (see his paper on 10 top mistakes in web design)

• The design must be related to a user's tasks

• A well-designed site downloads quickly

• A good design doesn't break browsers

• No gratuitous backgrounds or animation

• Should be internally consistent

• Should be externally consistent in use of navigation (Jakob Neilson's comments on need for standard navigational aids such as not changing colours of hypertext links)

• Variety in navigational aids (David Siegel's response)

53

Bad Web Design?

There was much disagreement over the Seigel "hack" illustrated

• Image on the next page stored in a 1 pixel by 1 pixel

• The image is downloaded while main image is being examined

• Moving to the next image it appears to load quickly

• "Neat trick" - David Seigel• "Hack", "future maintenance

problems", "no semantics"

<IMG SRC=image5.gif height=1 width=1>

image4.gif

54

Other Papers

The Proceedings contains papers on:

• Beyond HTML• Multidimensional Web Search• Searching Heterogeneous

Sources• Strategies for Resource

Location• Learning User Preferences• Usability Issues• Counting on the Web• Site Mapping and Syntactic

Analysis• Information Access• Automatic Interface

Generation

• Database Query Techniques• Caching • Embedded Systems• Highly Configurable Web

Services• Distributed Servers• Accessibility Using Audio• Real Time Video• Real Time Audio• Security and Payment

55

Conclusions

• The hit of the conference was XML• Java was the hit two years ago - now it's

accepted as the language of the Web • W3C are facilitating developments of web

protocols• Closed nature of W3C standardisation

creates uncertainty but seems needed following problems with over standardisation (e.g. HTML 3.0)

56

Challenges

The rapid development of the web poses various challenges for the community:

• Deploying new facilities in a timely manner• Choosing the winners• Spotting the winners• Minimising future web re-engineering

How can the community address these challenges:• Institutionally• Regionally• Nationally• Internationally