Transcript PowerPoint 97/2000 format

Electronic Magazines:
Issues in Implementation
Brian Kelly
UK Web Focus
UKOLN
University of Bath
Email: [email protected]
Acknowledgements to Bernadette
Daly, Editor of Exploit Interactive
(1998-9)
Aims:
• Describe approaches to
publishing a web magazine
aimed at European library /
information community
• Share experiences
• Describe other approaches
• Provide advice to those
wishing to publish web
magazines
UKOLN is funded by the Library and Information Commission, the Joint Information
Systems Committee (JISC) of the Higher Education Funding Councils, as well as by
project funding from the JISC and the European Union.
UKOLN also receives support from the University of Bath where it is based.
Contents
•
•
•
•
•
•
2
Background
The User's View
Searching
The Editor's View
Other Interesting Approaches
Conclusions
About UKOLN
UKOLN:
• UK Office for Library and Information Networking
• Based at University of Bath
Awareness and Advisory
Services
• UK Web Focus
• Interoperability Focus
• Public Library Networking
Focus
3
Research Interests
Metadata
Distributed libraries
UK and EU project work
Dissemination
Events (workshops,
seminars, etc.)
Publications
Web Magazines
Experiences
UKOLN publishes two web magazines, with a
third due to be launched shortly
Ariadne
• First published in January
1996
• 22 issues to date
• Originally funded by eLib
• Long-standing first
generation web magazine
4
Exploit Interactive
• First published in May
1999
• Four issues to date
• Funded by EU
Telematics For Libraries
Programme
Cultivate Interactive
• To be released in a few months
• Funded by EU DIGICULT
Programme (covers museums,
archives and library areas)
Ariadne
Ariadne:
• 22 issues published
Jan 1996 - Dec 1999
• Published every 3
(originally 2) months
• HTML pages created
using HTML
authoring tool
(HoTMetaL,
FrontPage, etc.)
• Apache web server
on Unix
5
http://www.ariadne.ac.uk/
Exploit Interactive
Exploit Interactive:
• 4 issues published
(every three months)
• Informs communities
on Telematics for
Library projects
• Aims to provide
quality content, be
innovative and be
standards
compliant
• NB: Little funding for
software
development!
6
http://www.exploit-lib.org/
Exploit Int. Architecture
7
Exploit Interactive:
• Runs on a Windows NT server
• Microsoft SiteServer software used
Why?
• Gain experiences of NT-based information system
in order to help in advisory role
• Make use of shrink-wrapped solutions to minimise
software development
• Investigate whether standards-based solutions can
be implemented in a Microsoft environment
Resources
• Funding for 0.5 FTE over 2 years
• Editor, project manager plus bought-in systems
support effort
D
The User View (1)
Exploit Interactive:
• Sub-menu pages for Features, Regular Columns, etc.
• Context-sensitive
navigational bars
• Links from all
pages to:
–Main entry point
–Issue entry point
–Editorial
–Features menu
–Regular Columns
menu
–News and Events
page
–Et Cetera menu
–Search facility
8
D
The User View (2)
Cascading Style Sheets
(CSS) used to:
• Define margin settings
• Define fonts
• Background for headings
Benefits of Cascading
Style Sheets:
• Ease of maintenance
(appearance defined in
single file)
• Re-use of resources
• Performance
• Accessibility
9
Behind The Scenes
Behind the Scenes:
• Article fragments used and assembled into HTML
resource using ASP (Active Server Pages)
• ASP server script used for server-side includes
(SSIs) and for manipulating variables
article_title="The CULIVATE project"
issue_num="1"
Assignment of variables
<title>Exploit issue <%=issue_num%>:
<%=article_title%></title>
…
<h3>Citation Details</h3>
<p><%=article_title%>, <%=author%>
Exploit Interactive <%= issue_num%>
10
Use of variables
The File Store
The file store structure for a typical article is shown below
issue4
cultivate
default.asp
assembles fragments and
processes variables
content.ssi
content – simple HTML
fragment
resources
citation details (based on
author and title variables)
citation.ssi
Some files stored next
to article
Others stored in central
area (parameter
passing)
11
article_defaults.ssi
author, title, etc. variables
figure-1.gif
The Bigger Picture
In order to minimise maintenance, we want to store scripts,
CSS files, etc. once and include them as appropriate
12
default.asp
<!--#inc file="article_defaults.ssi" -->
Read article variables
<!--#inc file="../issue_defaults.ssi" -->
Read issue variables
<!--#inc file="../../global_defaults.ssi" -->
Read global variables
<!--#inc file="/resources/ssi/html_dtd.ssi" -->
Define HTML DTD
<html><head>
<title>Exploit Interactive issue <%=issue_num%>:
<%=article_title%></title>
<!--#inc file="resources/ssi/dc_metadata.ssi" --> Create DC metadata
</head>
<!--#inc file="resources/ssi/nav_top.ssi" -->
Navigational bar
<!--#inc file="content.ssi" -->
Include content
<!--#inc file="citation.ssi" -->
Include citation
…
Problems - Style Sheets
Cascading Style Sheets (CSS):
• Separates content from presentation
• Reduces maintenance
• Simplifies HTML (no
complex tables,
clear.gif hacks, etc.)
13
• Accessible resources
But Netscape (esp.)
implements CSS very badly:
• Images can float over text
• Too much white space
around tables
• Problems in printing
• Other minor differences
Solution to CSS Problems
In order to address CSS problems:
• "User-agent" negotiation used to
include simpler CSS
file to Netscape users
• Uses a single SSI
fragment
• Tables stored in single
table cell – a hack 
• Should images be
included with
conditional SSI?
Also use CSS @media to
allow different style sheet
for printing
14
Searching
15
Exploit Interactive enables UKOLN to demonstrate its
research interest in metadata:
• Capabilities of SiteServer search facility investigated
• Fielded searches
provided. Can search
by Author, Title,
Abstract and full text
• Uses Dublin Core (DC)
metadata
• DC metadata (HTML
format) created using
ASP fragment
• Can easily create
metadata in other
formats (e.g. RDF)
Errors
The 404 error message has been tailored:
• Exploit Interactive look and feel
• Embedded search facility
• Tailored message:
The link to this resource
is incorrect – please
inform web master
or
You have entered an
incorrect URL
16
Extending the Web Site
The web site is being extended to provide more than simply
access to information
Experiment to search a project's
web site, from an article about
the project
17
Evaluation of discussion software to
enable readers to comment on articles
The Editor's View
Development work which may not be
noticeable by readers includes:
• Promotion of the web site
Making sure that people know about the magazine
and can easily find it
• Checking the web site
Ensuring that the web site works, is accessible and
conforms to appropriate standards
• Auditing the web site
Monitoring the popularity of the web site and being
pro-active in maximising popularity
18
Web Site Promotion
Promotion of web site addressed by:
• Having own domain name:
<www.exploit-lib.org>
Note shorter variants (e.g. exploit.org,
exploit.com, etc.) are not available.
• Having short, easily-cited URLs:
http://www.exploit-lib.org/issue1/pride/
Note use of server default file name to avoid longer
and possibly confusing file name and extension:
http://www.exploit-lib.org/issue1/pride/intro.htm
http://www.exploit-lib.org/issue1/pride/intro.html
http://www.exploit-lib.org/issue1/pride/Default.asp
19
Finding The Web Site
Aim:
• We want web site and articles to be indexed by
search engines and catalogued by directories
Approaches:
• Articles contain "AltaVista" metadata (generated as
described for DC metadata)
• Internal links to directory, avoiding links to .asp
extensions
• Use of robots.txt file and <meta
name="robots" content="noindex"> tag to
avoid indexing of unnecessary resources (e.g.
statistics, etc.)
20
Submission to Search
Engines
Exploit Interactive resources submitted to search engines –
see <http://www.exploit-lib.org/issue4/
promotion/>
1. New issue URL submitted
to AltaVista
21
2. Evaluation of Submission
software
3. Creation of single page
containing links to all articles
Does It Work?
Exploit Interactive articles:
• First (out of 421,122) for "exploit"
in AllTheWeb
• Fourth for "news exploit"in Google
• Fourth for "(electronic and paper)
near delivery" in AltaVista
22
Note: Searches found from referrer logs
Notification of a New Issue
Notification of new issues is sent to several mailing
lists. But what about readers who aren't on the
mailing lists?
http://www.exploit-lib.org/notify/
We now provide a
notification service
of new issues using
Netmind
23
Checking The Web Site
Prior to release of new issue we use:
• SiteServer Analysis to check broken links
• Bobby software to check accessibility of web site
24
Summary
On 10 January 1999:
906 pages
5,180 internal links
1,818 external links
0 broken links
Web site passes Bobby
accessibility test
Broken Links – Whose Problem?
Internal links are responsibility of editors
But what should be done with broken links in articles?
• Fixed: the web master is responsible for providing
a quality web site
• Nothing: articles should be left as is
• Flagged (and perhaps fixed): giving reader an
indication of original status
Our approach:
25
• Authors provide links as
references with visible URLs
• External links indicated by icon
• Icon and link change if link is
broken
• Display external links in new
window (personalised option?)
Checking The Web Site (2)
In issue 4 we provided a service for the readers to
check for and enhance access to articles:
• Translation Service
• HTML Validation
• Accessibility Check
• …
The link to further
information is stored
as a single SSI file,
allowing changes to
be made just once
26
Auditing The Web Site
We use SiteServer Analysis to provide statistics
In issue 4 evaluated third-party services which provide (e.g):
• Last 10 visitors to page
• Profile over time
• Referrer for last 25 visitors
Third-party Service Issues:
• Performance
• Dependency on 3rd party
• Open vs. restricted access
to statistics
NOTE: Comparison of Nedstat and
SiteMeter can be easily made due
to use of SSI fragments
27
Who's Linking To Us
The LinkPopularity.com web
site is used to:
http://www.linkpopularity.com/
• Provide information
on numbers of links
to web site
• Monitor growth in
nos. of links
• Identify friends
• Spot gaps
Information on links to web
site and nos. of resources
indexed is useful and is
28being monitored regularly
The eLib Programme
eLib programme funded several
e-journals – see
<http://www.ukoln.ac.uk/
services/elib/projects/>
http://elj.warwick.ac.uk/
http://www.ch.ic.ac.uk/clic/
Much can be learnt from
the experiences of the
eLib projects
29
Other e-journals: EPRESS
• eLib-funded project which provides an administrative
system (rather than a publishing one) for refereed
academic journals
• Addresses (time-consuming) administrative issues
• Used by Sociological
Research Online and
Journal of Artificial
Societies and Social
Simulation
See <http://www.epress.ac.uk/>
30
Other e-journals: D-Lib
D-Lib magazine:
• Currently mirrored at 4 locations including UKOLN
• Is making use of DOIs (Document Object Identifiers)
• Each article has DOI which provides persistent,
location-independent identifier
• "Handle" system can return multiple values e.g.
10.1045/december99-gladney can return
five URLs
• Interesting idea for the future (resolution by proxy or
browser plugin)
• Use of multiple values could be used for related
resources (e.g. metadata, translation, etc.)
• Exploit Interactive is considering a similar approach
31
Other e-journals
Library and Information Studies departments are active in
producing e-magazines and exploring various issues
For example see Information Research, published by the
Department of
Information Studies,
Sheffield Univ.
32
http://www.shef.ac.uk/uni/academic/
I-M/is/publications/infres/ircont.html
Exploit Interactive's Key Features
Exploit Interactive uses (normalised) resource
fragments for:
• Server-side scripts
• Navigational elements
• Standard text fragments
• CSS
Separation of Content / Navigation / Functionality is
useful for:
• Ease of maintenance
• Introduction of new functionality (e.g. print all
articles, personalisation)
33
Challenges
Use of resource fragments:
• Requires advance planning
• Some complications in use of authoring tools
• Some complications in validating resource
fragments (no compilers to spot, say, unassigned
variables)
• Useful to have a programmer's or DBA perspective
(identification of reusable fragments)
In addition use of ASP variables / file-based fragments
means:
• NT-dependencies (until Open ASP matures)
• Difficulties in managing variables which are defined
in scripts
34
Future Developments
Make use of article and navigation fragments to provide a
configurable interface
http://www.ljdigital.com/
Navigational bar:
Located:
• Top and bottom
• Side (in frame
or embedded)
Configurable:
• Sections of interest
• cf MS Office 2000
References & Related Links
As above
35
Future Developments
Ideas for the future:
• Variables stored in database (NB: performance issues)
• Use of XML
For Metadata and Articles (e.g. news items)
Would also like to provide
enhanced searching
(structured browsing):
• Search issue 1 & issue 3
• Search Regular Columns
• Search Regular Columns
in issue 1
• Search for articles about
EU projects
• Counts nos. of articles in
these categories
36
Future Developments
Processing articles is time-consuming.
Would like to automate processes:
Enter document
Browse
c:/docs/article.doc
Document Type
•
MS Word
HTML
Abstract
Keywords
TAP, digitisation, …
Programme:
37
•
ELIB
EU (5th Framework)
EU (4th Framework)
• Replace email
submission by
web form
• Automate
conversion to
HTML
• Author submission
of keywords
metadata
• Automated upload
to draft area for
editor review
Conclusions
To conclude:
• Exploit Interactive has shown the benefits of an
architecture based on:
– Resource fragments
– Style Sheets
– Server-side scripting
– Resources in neutral format to enable reuse
• It is possible to provide an innovative, standardsbased solution on an NT platform
• Similar solutions also possible on Unix (e.g. PHP)
• We look forward to further developments based on
use of database technologies / XML
Questions welcome
38