The Internal Landscape of a Distributed Website

Spider Redgold[HREF1], Information Technology Services, Griffith University, Nathan, Queensland, 4111, Australia. S.Redgold@mailbox.gu.edu.au


Abstract

Web teams are the 'maids of the universe', expected to keep beauty and order in a world where authentication is rarely from a managed and fully populated directory. Executive management want branding while business units, particularly ones that were early implementers of web communication, want independence with every on-line task directly linked from the front page. Every one thinks their site within the organisation should be higher, more obvious and more interactive than it is currently, and everyone knows someone who will do their site better cheaper or brighter than the current one. Web team members and managers struggle through a morass of conflicting needs; like swimming in baked beans. Current challenges for a web manager are o-dentification, content management, timeliness, multi-directional communication and the reduction of production costs. The internal landscape is multidimensional and using a multifaceted approach assists solutions delivery. This paper will use the Web Cube[HREF2] to introduce a multifaceted conceptual organization of site management.


Introduction

This paper addresses solutions that have been developed from the daily work of managing a large and diverse site with internal and external users, a distributed management model and organisational style standards. The issues in the foreground are those relating to large site management rather than those of web development. As Oliver and Johnson[HREF3] pointed out in " Growing Websites" Ausweb99, "Many websites do not address the issues arising from continued growth, are not managed beyond the initial phase of enthusiastic growth and are so difficult to maintain that further activity ceases." This paper focuses on managing a mature and distributed web site in organisations that are committed to web enabling of business processes, corporate information delivery and have core business intermediated via the web. Firstly the paper briefly addresses content management issues such as o-dentification (defined below), timeliness and communication, introduces a site management model, the Web Cube, which assists in addressing the named issues in a multidimensional manner. Then uses, as an example for discussion, solutions that are currently implemented. Griffith University, used as the example, has chosen to deliver its site through distributed content production, distributed independent web development and centralised site management.

Issues in Site management

O-dentification

"Identity is the crisis can't you see" (Poly Styrene, 1978, Virgin Records). O-dentity is on-line identity. Site visitors have on-line identity, content producers have on-line identity, as do content authorisers and also the documents themselves. O-dentification accrues at the moment of achieving a match between one's unique identifier and the materials, portal or interactive interface that is being accessed at the authority attached to the role linked with the unique identifier.

"Every person must be identified, allowed access to on-line materials based on their authority, within the life cycle of a document. Every document must be identified, to ascertain which person has authority to authorise the set(s) of users allowed access to the document, to update the document and to subsequently authorise it as correct and appropriate for publication according to its security/privacy category." (Redgold, S. 2000)

The data matching that must occur when a person needs to be identified, and allowed access to on-line materials based on their authority is different to that which must occur to manage the life cycle of a document. The latter requires the identification of a document, and the users authorised to see the document, the date that the document is due for review and update, the individual responsible for defining security/privacy category for publication. Many organisations do not have a single complete directory for the organisation. There are frequently multiple lists of persons, and even less comprehensive information or registries of documents. The current buzz is for comprehensive directory services. But many web teams have to meet this challenge by other means. Web teams have a history of providing enhancements that have extended the functionality of legacy systems. The incredible creativity of web workers driven by client demand has meant that institutions may find that the web security schema provides a good starting point for a meta-directory architecture. This schema has probably been pieced together by a variety of users, each responsible for designing security for their own areas; using the web to extend beyond the limitations of their business applications and limiting access to content using the security solutions made available to them by the web sysadmin.

Content Management

Content management is an area that is often confused with content production. Many sites rely on redevelopment for their content management. Iterations of a site are triggered by the need to update content and results in a new visual design. Clumsy work processes of emailing updates and changes to unknown beings in other departments, or at web development companies and internet service providers where the sites are hosted, can accompany content maintenance processes and contribute to poorly maintained content. Content management procedures are independent of the design and production of the visual elements of the site. The scope of content management includes: scaling of site growth to ensure adequate disk space and CPU cycles are available on high-speed disks for optimal delivery; data movement from development, to staging, to production environments; providing an audit trail of the authenticity and authority to publish; enterprise wide strategies for content classification; provision of templates to minimise costs of content production and maintain style requirements; template testing for browser and operating systems independence; procedures to ensure that all of these are carried out in an effective manner and reporting on site usage to content providers. If the site is required to be accessible to legacy browsers, site visitors who use disability access aids and low-end technology then this category of content management can also include provision of alternatives for legacy and low end technology and testing for the popular screen to voice readers and possibly language translation applications.

Timeliness

Planned timeliness includes the planning and implementation of cycles of audit, update and review and signoff of correctness and perhaps quality of content archiving and versioning notifications and communications between content providers, developers and web maintenance teams. How often do you visit a site where the 'last updated' insert is clearly at odds with the content of the page. The updater has added the required edit without considering the full content of the page for correctness, indicating under-resourcing or lack of ownership of quality. The last updated tag is not automated or controlled to link to the system date of the file. Reporting on site usage, site integrity and the methods for delivering reports are included here as well as within the content management category as these reports contain valuable strategic communication that impacts on timeliness. Appropriate timeliness initiatives by the web management team can ensure that initiatives such as a marketing push by the organisation is accompanied by appropriate content on the web. A habit of regular update and audit of site content can create awareness in unit managers so that site content becomes part of change management.

Communication

The matrix of communication for web site management references the content of the web but is not the content of the web. In some cases, there is web content that is part of the site management communication, for example, 'The Expediter"[HREF4] where the site is specifically used to communicate site information and procedures to the content producers.

Interactive version of Matrix
  Figure 1: Communication Matrix - Image link[HREF5]

The communication occurs between web team and users areas of diverse technical responsibility, content providers, content custodians, developers, policy makers and senior management. The communication is about technology and site management issues, policy and procedures about the state of the content in relation to site management issues, rather than about the meaning of the content contained with the documents and web pages. The effectiveness of communications between the web team and senior decision makers can ensure that issues arising from new technologies and changed position responsibilities are assimilated into HR policies and infrastructure and acquisition schedules. Communication can be to negotiate outcomes, gather information, to report on site metrics and to inform on policies, procedures and options.

The Internal Landscape of a Website

interactive model of web management cube
Figure 2: The Web Cube - Image linked to[HREF2]

The web cube is a tool to focus on the issues relating to large site management rather than those of content development. It describes topography for the internal landscape of a site and can be used for any site. The centre of the cube is investigated from each surface, asking the questions relevant to a robust regime for site management and maintenance. Three faces of the cube reflect the tools, user experience and ecology of the site, two faces are communications and the solutions, while the base of the cube, the final face, is formed from the questions within the unique context of each site environment. Each face comprises 3 slabs, which are further divided into 3, producing 27 cubes and making the whole.

The Example Site

Griffith University completely redeveloped its web during 1998, and has distributed web development and content development with centralised site and content management. There is a web site policy [HREF6]. This policy is reflected in and supported by a vision for access to electronic resources internal document.

Diagram of Griffith University Schema for Access to Electronic Resources
Figure 4: A Vision for Access to Electronic Resources at Griffith University

The group of people responsible for the site are called webhumans. The web team, called the IntRnet team as a contraction of both internet and intranet, is a subset of that group and its responsibilities at Griffith University include management of the site, support of the procedures that ensure accuracy, timeliness and future proofing of the site. The web manager and team are not responsible for most of the web page development or the content production of the site. This is the responsibility of the wider group of webhumans. There is a three-ply strategy for the user experience of this web site. The site strategy is woven from the navigation, the look & feel and the content management, which comprise the front face of the cube. This strategy is supported by a variety of technologies and business processes that include scripts, automation and templates, metadata and applications, which are the slabs comprising the Tools face of the cube. These are defined and guided by policy and must be attainable by available resources, and integrated with existing architecture (and infrastructure), the Ecology face.

The communication face connects through the centre of the cube to solutions and tools via the footer of every page. The footer, added via a server side include (tool) contains a hyperlink toa cgi-script that triggers a form. This allows a site visitor to send feedback or ask a question. The output of this form (communication) is delivered to the individual responsible for publishing the content of that page. Feedback is received at this point and is dealt with within the unit responsible for the site or, if necessary, is directed to a more appropriate contact to deal with the feedback. There are approximately 220 in the full group of webhumans. They are content providers, the web developers and the content custodians for Griffith University. The IntRnet team has 4 staff and is only a small part of the full set of webhumans. It is the webhumans that are the real workforce and drivers of the quality and high standards of the GU web. There is at least onefor every business unit in the university. Webhumans are the key players in the success of a distributed content management. This distributed model is managed through a site that communicates policy, delivers templates and provides interactive discussion spaces for expediting the publication of GU On-line "The Expediter"[HREF4]. This site is the portal for webhumans giving them access to tools and documentation.

 

O-dentity

O-dentification within the university site is achieved with a mixture of in-house and shrink-wrapped tools. These solutions ensure that the material is secure and accessible to defined sets of viewers. The specific use of the meta-tags "RightsUse" and "Prerequisites" communicates the o-dentification definition of the page access rights, Metadata templates and Metamaker[HREF7] are tools to support o-dentification, defined by the policy (ecology), UNIX access control lists (ACLs) to maintain publishing access secure rights via file transfer protocol (FTP) to a staging server. These tools manage the o-dentification of custodians and publishers of the page. The Netscape Directory Server, implemented July 1999,handles the server authentication processes. Prior to the implementation of the Netscape directory server the university used the native security provided by Netscape web server application. Any or all of the following currently controls o-dentity and o-dentification: IP control; password control by individual; by role/group and some legacy security to the branching file structure using .nsconfig files A secure portal using ACLs to manage publishing, maintenance and deletion of pages and files, PERL scripts and an IIS served database directs web visitor communication direct to the webhuman responsible for the content of each page. On-line messaging system provides ability to identify visitors by IP and the site they visit to deliver specific messages at monitored times and frequency. Oracle web applications deliver secure financial reporting and student access to enrollment details and results.

Content Management

The IntRnet team contributes to pre-production by making consulting services available to business units free of charge to assist each unit's webhuman and content custodian in conceptualising the business unit site and the documenting their information needs. Each business unit is responsible for content production and can have this done within their unit, by external developers, or by individual arrangement by any other unit in the university. The web policies and standards apply regardless of content production choices. The IntRnet services team is responsible for providing tools and support for post-production, quality control, version control and archiving. Templates provide a mix and match set of templates that maintain the University style while allowing elements to have distinctive sites.

Future-proofing of the Content

The frameset style has been designed to allow a complete change of look and feel with minimal work. When the University chooses a new look and feel, primarily navigation pages and graphics elements will require changing. There is a naming convention in place to allow batch processing of many of these changes. Minimal content changes will be required. There is also a naming convention for DNS entries for web servers that separate web site references from machine names. This ensures maximum portability of the web sites that are served from a variety of machines. In a recent disaster recovery exercise, one web site, consisting of more than 140,000 files was taken off-line and replaced by a staging replacement site in less than 35 minutes.

Tools to Manage Content

Timeliness

Tools to Manage Timeliness

One of the key deliverables set in 1998 was to produce a site that would not have out of date content. No content page can be published without a review and expiry date set in the metadata. Webhumans control and set their own review and expiry dates. The maximum time between reviews is eleven months unless the site is archived. When a site is archived it becomes marked as archival material and cannot be edited. Reports are produced weekly for webhumans and include site integrity, site usage, expiry and review of files. These are published to the web and divided into single business unit granularity. Notifications of expired files are sent by email and linked to the full expiry and review report on the web reports site.

Communication

The success of this strategy and model is measured by the effectiveness of the communication. The GU web procedures and policies were implemented in an 8 month schedule and involved significant change management and training. Maintaining communication remains a significant resource overhead. Routine visits and discussion meetings are held for webhumans on each campus. Presentations are given to senior and general staff meetings and regular meetings are held with the training group to ensure that relevant and accurate information is being given through the training sessions. Other communication tools used include:

These tools are supported by documents including policy explanations, help files and templates for the mix and match standards that complete the current GU look and feel.

Conclusions

The University in 1997 had a site that was untended and many pages were abandoned and incorrect. The decision to radically change the web management structures has been a high profile project that was often contentious.

There has been a noticeable acceptance as individuals become more familiar with the distributed control model, collaborative communication and negotiating between business units to ensure the whole site meets the needs of the community. The policies and responsibilities are clear and the improvement cycle is working. A measure of this success is that each week approximately 5% of the site is due for review and content providers and developers are being creative while still working within the standards.

A review of university web content management by the Pandora Project of the National Library of Australia has indicated that Griffith University is a leading site for web content management and hence has been mooted to take part in a pilot for tagging academic research output as 'of enduring merit'. There is a relatively high cost in training the individuals who are the webhumans for each unit. Frequently these are administrative staff with little or no experience with web development tools. The importance of ensuring that each unit is responsible for its site and content must be considered in relation to the cost of training webhumans to maintain their content. If the model is changed, any new model for web development will still include content control and production by the individual unit, even if web development responsibilities are changed. The experience of Griffith University through a complete redevelopment has shown that effective content management, supported by policies and an information architecture, will ensure that an organisation does not fail in its commitment to use the web as a major communication tool with the internal and external community and the avenue of access for current and future business functions.

References

Oliver, J., and Johnson, D. (1999). "Growing Websites",[HREF3] Ausweb99 Proceedings of the Fifth Australian World Wide Web Conference, Ballina, NSW

Styrene, P. (1978). X-Ray Specs, Lyric from "Identity" Virgin Records

Redgold, S. (2000). The Web Cube, WWW9 - Proceedings of the Ninth International World Wide Web Conference[HREF8]. Amsterdam, NL: Foretec.

Hypertext References

HREF1
http://www.gu.edu.au/ppages/S_Redgold/
HREF2
http://www.shoalhaven.net.au/~cyberzonne/webcubemodel.html)
HREF3
http://ausweb.scu.edu.au/aw99/papers/oliver/paper.html
HREF4
http://www.gu.edu.au/expediter
HREF5
http://brainserver.thebrain.com/get.asp?i=5f98c
HREF
http://www.shoalhaven.net.au/%7Ecyberzonne/webcubemodel.html
HREF7
http://www.gu.edu.au/expediter/meta
HREF8
http://www9.org/


Copyright

Spider Redgold, © 2000. The author assigns to Southern Cross University and other educational and non-profit institutions a non-exclusive licence to use this document for personal use and in courses of instruction provided that the article is used in full and this copyright statement is reproduced. The author also grants a non-exclusive licence to Southern Cross University to publish this document in full on the World Wide Web and on CD-ROM and in printed form with the conference papers and for the document to be published on mirrors on the World Wide Web.


Proceedings ]


AusWeb2K, the Sixth Australian World Wide Web Conference, Rihga Colonial Club Resort, Cairns, 12-17 June 2000 Contact: Norsearch Conference Services +61 2 66 20 3932 (from outside Australia) (02) 6620 3932 (from inside Australia) Fax (02) 6622 1954