Transcript Powerpoint 97 Format

Future Of The Web
Brian Kelly, UK Web Focus
UKOLN
University of Bath
Bath, BA2 7AY
http://www.ukoln.ac.uk/
1
UKOLN is funded by the British Library Research and Innovation Centre, the Joint
Information Systems Committee of the Higher Education Funding Councils, as well
as by project funding from the JISC’s Electronic Libraries Programme and the
European Union. UKOLN also receives support from the University of Bath where it
is based.
Contents
Data Formats
• HML 4.0
• CSS 2.0
• Dynamic HTML
• XML
Transport
• HTTP/1.1
Addressing
• URLs
• URNs
• DOIs
Metadata
• PICS
• RDF
Implications for the Community
2
Web Protocols
The web was originally
based on 3 protocols:
• Data Format
(user interface):
HTML
• Transport:
HTTP
• Addressing:
URLs
3
Data
format Transport
Addressing
HTML
HTML:
•
•
•
•
Defines structure
SGML application
Platform and application independent
Simple and therefore easy to create
but:
• Netscape / Microsoft browser wars began
• HTML extensions appeared
• (Proprietary) scripting languages add new
functionality and also complexity
• Universal access begins to be lost
4
HTML Develops
5
During 1997 recognition of dangers of HTML
wars
Netscape and Microsoft began working with
W3C on development of new protocols
HTML 4.0
• W3C recommendation
• Primarily defines structure
• Layout tags for backwards compatibility
CSS 2.0
• W3C draft
• Used to define appearance
CSS 1
CSS level 1:
• W3C recommendation in December 1996
• Partly supported in IE 3.0
• Early designs attempts confirmed power of CSS
but highlighted backwards compatibility issues
6
CSS 2
CSS level 2:
• Provides richer functionality
• Addressed backwards compatibility issues
7
http://www1.shore.net/~straub/
wprmultb.htm
Making HTML Dynamic
DOM:
• Document Object Model
• W3C activity
• Enables client-side scripting languages to
manipulate HTML and CSS elements and
their content
• Sometimes referred to as dynamic HTML
But:
• Netscape's implementation is nonstandard
8
DHTML Example - 1
Making documents "collapsible" is very easy:
on mouseclick in heading
set visibility and display of body off
9
DHTML Example - 2
Dynamic HTML can be used to develop improved
user interfaces.
In this example an expanded menu is provided.
10
DHTML Example - 3
DHTML can be used to provide interactive
systems - e.g. for CBL, simulation
In this example the
pieces can be moved.
11
Compatibility Issues
Compatibility with old browsers / other
browsers can be provided by:
• Using safe, backwards
compatible features
• Checking for browser
type by client-side
scripting language
• Checking for browser
type by server
Alternatively ignore backwards compatibility
12
Using DHTML Locally
Can create a variety of dynamic sitemaps
Your website
13
Remote websites
Your pointers
Information Landscapes
Developments such as Web Collections will
enable metadata for groups of related
resources to be defined for:
• Providing enhanced user navigation
• Facilitating printing, off-line browsing, ...
• Facilitating indexing
• etc.
The term Information Landscapes has been
coined to refer to changing, configurable views
of networked resources
14
Information Landscapes
Landscapes can be developed by / for:
• Institutions
• Service providers
• Groups (e.g. undergrads)
• ...
Information Landscapes can be used for
delivering information based on:
• Institutional needs and interests
• End user's technologies / interface preferences
Implications:
15
• Ownership, legal and copyright issues
• Change control
• Visibility
• Maintenance
What demand is there?
The End for HTML?
HTML 4.0 and CSS 2.0 provide a rich
environment for describing document structure
and appearance
But:
• Standardisation of new elements is slow
• Supporting communities with specific needs
(e.g. Maths) will result in large, complex HTML
• Applications have specific needs (e.g.
<STUDENT-NUMBER>
• HTML 5, 6, … does not appear to resolve
these issues
16
XML
XML:
• Extensible Markup Language
• Can be regarded as a lightweight SGML
• Supported by SGML community
• Some support in Internet Explorer 4.0
• Netscape promising support in future
• XML 1.0 became a W3C recommendation
in Feb 1998
• See <URL:http://www.w3.org/XML/>
17
XML Applications
MML / CML
Scientific communities have defined Math and
Chemical Markup Languages
CDF
Channel Definition Format provides a proposed
standard for push technologies
RDF
Resource Description Framework provides a
proposed standard for a metadata architecture
...
18
CDF Example
NewsAgent:
• an eLib project
• evaluating potential
for CDF for pushing
information
Software such as
FrontPage 98 have
CDF wizards
19
http://newsagent.ukoln.ac.uk/channel/
<CHANNEL HREF="http://newsagent.ukoln.ac.uk/channel/top.cdf">
<LASTMOD="1998-03-05T00:39" />
<SCHEDULE>
<INTERVALTIME DAY="1"/>
</SCHEDULE >
<ITEM HREF="http://newsagent.ukoln.ac.uk/channel/Events.html">
<TITLE>Events</TITLE>
Typical CDF File
</ITEM>
Transport
HTTP 1.0
 Widely used
 Inefficient
 Cache unfriendly
HTTP/1.1
 Addresses deficiencies in HTTP/1.0
 Cache-aware
 Available now (e.g. latest version of
Apache)
 Supported in ~ 40% of UK HEIs
 Browser support?
20
Addressing
URLs:
• Provide addressing for the web
• Consider http://www.dec.com/internet/
applications/
• Problems:
– Company reorganises website
– Individual leaves
– Company is taken over!
21
Solutions (1)
URNs:
Uniform Resource Names: make identification of
Internet documents possible.
URN identifiers are persistent and unique.
See <URL: http://www.ietf.org/
html.charters/urn-charter.html>
PURLs:
Persistent URLs: Point to a resolution services
rather than the resource itself.
See <URL: http://purl.oclc.org/>
22
Solutions (2)
DOIs:
DOIs use CNRI's Handles
• Document Object Identifiers
• Costs $1,000 to register
http://hdl.handle.net/10.1016/
S1384107697000225
23
hdl://cnri.dlib/august95-contents
Addressing: Conclusions
Conclusions:
• No clear technological winner
• Performance issues
• Is it just for "publishers"
• Business model uncertain
• Backwards compatibility and "real
world citation" issues
• Still need for information providers to
give careful thought to URL naming
schemes
24
Metadata
Metadata:
• The missing architectural component of the Web
• Many developments in progress:
– PICS: "this resource contains bad language
rated as X"
– Dublin Core: the author's name is xxx and his
email address is yyy
25
– Web Collections: these set of resources are
related and should be download together /
printed as a single document / …
– Distributed Authoring: the following can
update this document
– ...
RDF
Resource Description Framework (RDF):
• Generic infrastructure for metadata
• Based on XML
• See <URL:http://www.w3.org/RDF/>
26
Using The Technologies
How / when are new technologies to be
deployed?
• Use "safe" features to ensure backwards
compatibility
• Check for browser functionality at server
• Check for browser functionality at client
• Deploy on Intranet for access to internal
resources / external resources
• Use tools which can regenerate systems when
standards change / develop
• Which tools?
• Need to avoid the "tyranny of the template"
27
Web Architecture
Web architecture is:
• Growing in
complexity
• Maturing,
becoming
more robust
Must understand
architecture to
avoid:
• Escalating costs
• Degradation of
service quality /
functionality
28
Data Transport
format
Address- Metadata
ing
Community Issues
Issues which the community need to address:
• Can we agree on a minimum recommended
browser spec?
• Can we use more sophisticated technologies to
cater for variety of browser technologies?
• How can we share experiences / knowledge of
what we're doing?
29
WebWatch
Nos. of hyperlinks
Project based at UKOLN
to monitor web
developments in UK
Analysed various
communities:
• University entry points
• eLib projects
• Public libraries
30
Server software
File size (HTML)
Web Focus ROADS Gateway
A ROADS gateway
(software used by SOSIG,
OMNI, etc. gateways)
being developed to
support UK HE web teams:
• AUPs
• Guidelines
• …
Your feedback
needed
31
Questions?
32