Report Back for Concertation Day

This document gives summary of the report back for the eLib Technical Issues Concertation day, held at the University of London Computing Centre on 8th November 1996. There were 35 participants at the day, plus three organisers - Chris Rusbridge, Kelly Russell and Brian Kelly.

Breakout Session 1

Four groups of approximately 10 people were asked to produce a list of important technical issues.

Group 1

• Cookies
• Server & client support for cookies / server implications for large user base
• Server and network performance & architecture
• Scaleability of services / mirroring and caching (including copyright implications)
• Access control
• Delivery formats - style sheets, on-the-fly generation of HTML
• Metadata
• Groupware - electronic conferencing, annotations, etc.
• Text summarisation

Group 2

• Authentication and access control
• Digitisation of text and images
• Database integration
• Charging mechanisms

Group 3

• Electronic journals - Acrobat issues
• Link maintenance and automated creation
• Network security
• Specification of end user system / user demands

Group 4

• Authentication - national and international services
• How to send documents - ease of publishing documents
• Adobe PDF issues (indexing, searching, performance, printing, adding value - metadata, links)

Breakout Session 2

In the light of the comments from breakout session 1 and discussions over lunch, the following four thematic BOFs (birds of a session) sessions were identified for breakout session 2:

1. Authentication
2. Document formats issues and user issues
3. System issues - cookies/ server architecture
4. Replication / scaleability / interoperability / metadata

Document Formats Issues & User Issues

• Adobe PDF:
  • Advantage - page representation
  • Problems with printing (e.g non Postcript drivers, performance)
  • Good for referencing / searching (but indexing tools cost e.g. Verity) / security / copyright
  • See http://www.adobe.com/acrobat/main.html
• HTML Tools:
  • Myrmidian (Mac) useful tool [Quark - HTML directly]
  • RTFTOHTML (see http://www.sunpack.com/RTF/rtftohtml_overview.html / Microsoft Internet Assistant (see http://www.microsoft.com/msword/internet/ia/
  • Frontier (Mac currently) - see http://www.scripting.com/frontier/
• Link Validators
• SGML - HTML conversion tools, such as:
  • OMNIMARK, CLIC and Cost
• eLib coordination
  • Need for lis-elib-techies Mailbase list
  • eLib FAQs
• Statistical Analysis
  • webstar logs
  • Frontier

Authentication

This is felt to be a difficult area

• Problem - maintenance of large nos. of user ids
  • Subscription (e.g. paid services. often done at HEI level)
  • Management of username and passwords
  • Security at point of access ("Is this request really from the student John Smith at Bath University?")
  • Data Protection
• Reports available:
  • No strong recommendations for solution
  • Solutions 2 years away (and still costly)
  • NISS and BIDS Athena project (note not DEC Athena)
  • Edinburgh, Manchester and Bath work - report available
• Recommendations for eLib projects:
  • Talk to NISS
  • Look at Common ID (e.g. extension of Midas solution)
  • Look at software such as Verisign
  • Look at cookie technology. Netskills could have a role to play in developing training materials about cookies (NOte see Tony McDonald's page at the URL: http://www.netskills.ac.uk/staff/mcdonald/how/cookies.html

System issues

• Database integration (e.g. CGI / Java / Perl to access Sybase, EFS) But expensive
  • Packages available e.g. IndexPlus (but unstable and buggy)
  • WWW gateways now bundled in with large DBMSs
• Electronic conferences (MUSHES, MOOs, MUDs)
• Video conferences - look at MBONE, CUSeeMe. PC clients not yet mature ( position now changing - CuSeeMe felt to be excellent on PC)
• Load on WWW servers - lack of tools for management, but stats availbale
• Apache is felt to be a good server
• Look at comparisons of Web servers (e.g. http://www.webcompare.com/ and the section on YAHOO! at http://www.yahoo.com/Computers_and_Internet/Software/Internet/World_Wide_Web/Servers/Comparisons/)
• Virtual WWW servers, multihoming
• Cookies
• Webstats - ERCOMS looking at - problems with stats
  • Analog and WebTrend packages
• What is a session?
• Some participants did not like proxies

Interoperability

• Need for eLib resources to be accessible from each other and searches to interoperate
  • 4 ways of doing this (see diagram)
  • Should be based on a standard. Two standards suggested for different layers - Dublin Core and Whois++
• Results
  • Processing of results
  • No standard metadata set - develop one?

Recommendations From The Day

• Let's use the lis-elib Mailbase list more
• Let's think about a lis-elib-techie list
• Look at use of hypernews, etc for eLib technical staff. (Brian Kelly (UK Web Focus) to coordinate small trial by eLib technical staff)
• Look at use of video conferencing software (e.g. MBONE). (Donal Hanna (Netskills) to coordinate small trial by eLib technical staff)
• Attempt to identify what we mean by the "virtual / electronic library". Put out requests for vision statements of what is meant by this term.
• Hold more events for eLib technical staff

Other Reports

• Andy Powell's report on security issues
• Clare McClean's report for Ariadne (Issue 6)

Brian Kelly, UK Web Focus, UKOLN, University of Bath