Search   |   Back Issues   |   Author Index   |   Title Index   |   Contents



D-Lib Magazine
May 2003

Volume 9 Number 5

ISSN 1082-9873

Developing a Content Management System-based Web Site


Clare Rogers
National Trust

John Kirriemuir

Red Line


1. Introduction

This article describes the development of a content management system(CMS)-based web site for the Joint Information Systems Committee (JISC) [1], a UK strategic advisory body. After introducing the JISC, some of the motivations for developing a new web site are outlined, as well as an overview of the site mechanisms and some of the development issues. Navigational functions are described, followed by a more detailed examination of one section, projects and programmes. We examine some of the issues involving training and supporting CMS users, and conclude with observations on the future development of the web site.

2. The JISC

The JISC is an independent advisory body that supports further and higher education by providing strategic guidance, advice and opportunities to use information and communications technology (ICT) to support teaching, learning, research and administration.

The JISC is financed by the UK further and higher education funding councils, and operates through a committee system, the membership of which comprises senior managers, academics and technology experts working in education. These committees are supported by the JISC executive, facilitating policy definition and the management of JISC-funded services and programmes of activities and projects.

The JISC executive comprises approximately 50 staff, distributed around UK educational institutions. As well as funded projects and services, a number of external consultants and advisors support content preparation, production and dissemination activities.

3. The JISC web site

In 1999, JISC Assist (the communications and outreach division of the JISC) embarked on a programme of activities associated with developing a coherent message to the community about the range of JISC activities. The JISC web site is the key component in developing this message and is at the heart of the JISC's emerging communications and dissemination strategies.

The need to increase community access to JISC information and resources in a coherent and coordinated way was central to the work. A primary directive of the redevelopment was to reduce the need for the community to understand the complexity of the JISC in order to identify useful resources.

A phased plan of activities for repositioning the JISC web site began with a period of research including three pieces of work [2]:

  1. A usability study
  1. An accessibility study
  1. A scoping study for a Projects and Outputs Database (POD)

The findings of the reports were complementary, and all pointed to the need for the JISC to reposition its web site to become a leading example of best practice in information management.

The usability report gave a number of recommendations for the short-, medium- and long-term. The short-term and some of the medium-term recommendations were implemented immediately in the first phase of web redevelopment, along with the recommendations from the accessibility audit. This included:

  • a more advanced search facility
  • a new look and feel
  • a site map giving a schematic representation of the site, and 'Help' pages
  • section and sub-section navigation
  • some site metadata
  • online booking forms

Design and standard navigation features were applied to every page by the web server, using templates to ensure a consistent look and feel. This stage of the development did not attempt to deal with the longer-term recommendations of the reports, namely to develop database technology to run the site and to implement content management systems. These technologies were still not widely adopted in educational web sites at that time.

Implementing the remaining recommendations required an extensive programme of reorganisation work, and in January 2001 JISC Assist proceeded with an OJEC tender for the second phase, redeveloping the web site to include workflow processes for information management and preservation issues.

Previously, JISC web content was produced by a small number of staff, using a centralised model to mount content and make changes. This approach often resulted in internal delays and frustration. The site lacked any automated features to maintain time-limited content such as news and events and, as such, this type of content could not easily be managed.

In 2001, a survey was carried out amongst typical users of the old web site, focusing on how easily (or otherwise) users found information for which they were looking. One hundred and thirty-one responses were received, revealing a number of problems with the site. Many of the comments focused on the difficulty of finding information. This was partly due to the fact that, as more content was added to the web site, it became more difficult to locate current information, by either searching or browsing, among the older content. Information was also presented under various JISC initiatives, requiring users to understand JISC structures in order to locate it. Further, a clear need emerged to address the issues of redundancy, customised information and aspects of the JISC's investment that were not well described.

In recent years, web development has changed. Advances in browser functionality now offer the developer more sophisticated methods for presenting information. However, there is a far greater emphasis on ensuring that this information is easy to obtain from the web site, with legal requirements for web sites to adhere to accessibility standards [3]. As the Internet has moved from being a novelty to a niche, and finally to a mainstream dissemination channel, so expectations regarding the quality and accuracy of its content, especially that of "trusted" organisations such as educational funding bodies, has increased.

The new web site would deliver:

  • Adaptivity, enabling the site to adapt to the needs of its users. This meant the site would change the information it presents to users based on a profile of their interests or sector. The information contained in the profile could be gathered either implicitly or explicitly.
  • Fast and efficient full text searching based on keywords, topic and thesaurus-based recommendations, detailed advanced searching.
  • New design with full navigational functionality.
  • Content management system - a simple web-based interface for creating and publishing the information. This includes defining users and groups and restricted access to the site for editing purposes.
  • JISC internal facilities: secure logon, and an internal (private) file area for storing/drafting web content.
  • Metadata, based on standards such as Dublin Core, and keyword descriptions to augment the searching and thesaurus functions.
  • Online forms for booking events and ordering publications.
  • A knowledge management and archiving facility.
  • Potential for further development of facilities such as newsfeeds, tracking systems, detailed web statistics and so on.

4. The CMS

4.1 Overview of development

A Content Management System (CMS) was developed [4] to allow the storage and manipulation of web site content. CMS users were provided with direct editorial access to defined and appropriate portions of the content. Such a system therefore allows distributed JISC staff, and other CMS users, to mount content on the web site as soon as it has been created.

Initially working to a formal specification, external contractors [5] developed the CMS. An online bulletin board facility, the Change Control System (CCS), was used by the developers, the JISC web manager (Clare Rogers) and an external consultant (John Kirriemuir). The CCS allowed error, enhancement or functionality requests, as well as queries to be logged, time-stamped and answered. By April 2003, nearly 300 CCS threads had been logged.

Image showing a typical CCS thread

Figure 1. A typical CCS development thread

The CMS uses ActiveX controls, operating through the Internet Explorer browser. Installation automatically takes place when a user logs on to the CMS for the first time on a particular PC; installation is only necessary once on each computer. Interaction with the CMS is an online activity (through modem or network), with the CMS based on a server at Eduserv [6].

CMS users can therefore manipulate content in the CMS from a number of locations. For example:

  • From their desktop PC at work or any other PC where the CMS has been installed.
  • ActiveX restrictions notwithstanding, from a PC in a public location such as a cybercafe or public library.
  • From a laptop. For example, a CMS user could attend a meeting, receive or create content in electronic format (e.g., meeting minutes), and then enter it into the CMS from either the meeting room (via network) or their hotel room (via modem).

To date the CMS has been used from various locations including: cybercafes, public library PCs, Internet kiosks, hotel Internet facilities, and even the middle of a loch using a laptop connected to a mobile phone. To test the "anywhere, anytime" nature of CMS access, users have also successfully installed the CMS and created content in several countries other than the UK (Belgium, Sweden, USA, Ireland, Canada).

4.2 CMS users

Each file in the CMS corresponds to a page on the JISC web site. There are three types of file permission:

  1. Browse: users can see and read the contents of the file in the CMS.
  1. Edit: users can edit the contents of the file.
  1. Publish: users can make the file publicly viewable on the JISC web site.

There are three kinds of CMS users:

  1. Normal users: these have access to a subset of the files available in the CMS. For example, the collections team group of users can see (browse), edit and publish files in the collections area of the CMS, but cannot see files associated with projects and programmes.
  1. Supereditors: these have automatic browse, edit and publish access to all files in the CMS.
  1. Superusers: these have the same access as supereditors, but can also manage CMS user accounts, as well as configure certain settings within the CMS. For example, superusers can edit the style sheet to which all pages on the web site are dependant for formatting instruction. Superuser accounts are strictly limited due to the potential for catastrophic deletion of directories and files, for security and for preservation of standards and appropriate use of the CMS.

CMS users are organized into groups. Each user has an individual username and password, while each group is assigned a set of access permissions by the superusers.

4.3 Using the CMS

When a CMS user logs on, two main windows are presented: the document tree and the editing window.

4.3.1 Document Tree

In the document tree window, the user can see the folders/directories and files to which they have access within the CMS. The directory structure and icons are very similar to those used in Windows Explorer. To add a new file, the user clicks on the appropriate folder and selects add from the list of functions at the bottom of the window. Other functions offer the ability to edit a file, publish it (i.e. make it visible on the web site), delete it (from the CMS), or suspend it (i.e., remove it from the web site, but retain within the CMS for future editing and publishing).

Image of the document tree

Figure 2. The document tree for a CMS user

4.3.2 Editing window

The editing window is used to enter and manipulate content within a file. For most files, this consists of two parts:

  1. Content editing pane: This is an adapted version of the Ektron eWebEditPro tool [7]. Through this pane, CMS users can edit content in both WYSIWYG and HTML modes. Many of the icons have functions identical to their Word counterparts (e.g., create bullet points or right-indent a paragraph). Additional functionality allows the CMS user to add styles that are dependant on the web site cascading style sheet.
  1. Data and metadata fields: The fields provided for metadata inclusion depend on the folder in which the associated file is stored. For example, a file associated with a project description will include a data field for specifying the associated programme (see section 6), while a file associated with an event will have fields for the start and end dates of the event. Some data fields are used when automatically generating web pages, e.g., current events; others are used to generate metadata for the associated web page to assist in indexing by external search engines such as Google™.

Image of the editing tool in use

Figure 3. The editing tool in use

Each file has an audit trail, which can be viewed by CMS users with browse access. Every time a change is made — such as an edit or publication — the date, type of change and name of the CMS user who made the change is added to the trail. Where a page is accessible by multiple CMS users, the audit trail enables the individual who made particular changes to be identified.

Image showing an audit trail

Figure 4. Example of audit trail

An important feature of the audit trail is its provision of access to previous versions of a file. This is useful when, for example, a CMS user deletes some important content from a file. When the mistake comes to light, a previous version can be re-published, restoring the important content.

4.4 Web and CMS design issues

4.4.1 Browser compatibility

One issue that came to light during the design and testing of the web site is the wide range of browser types, and versions of browsers, used in UK Further and Higher Education (FE and HE). by staff and students across educational institutions. In many cases, their choice of browser is restricted to that prescribed by the institution.

Considerable work was undertaken to ensure that as wide a range of browsers as possible could correctly display content from the JISC web site. For example, it was discovered that some UK universities still used earlier versions of Netscape (v 4.x), necessitating development and testing to ensure that users of this browser version could view the web site correctly.

4.4.2 Text-only viewing

A version of the web site with all graphics and images removed was deemed necessary for legislative reasons and so that users of text-oriented browsers (such as speech-output browsers) could access content.

The web site incorporates the free Betsie (BBC Education Text to Speech Internet Enhancer) parser [8], offering a link from every page to a text-only version of the site. Using this system, the user can also configure certain aspects of the display, such as the text size and colour, and the background colour.

This facility has proved highly popular. Surprisingly, feedback has indicated that some users, despite having access to the newest browsers and fast download connections, prefer to use this text-only version for its simplicity and speed of use.

4.4.3 Pictures and graphics

In keeping with the web sites of many other educational and funding bodies, pictures and images were kept to a minimum. While an abundance of pictures may make the site "prettier", they adversely affect users on slow modem connections, those who use text-only or disability browsers, or users who, out of choice, have images and graphics turned off.

A repository allows CMS users to both upload images into a shared internal area and to embed images from this area into their web pages. The CMS user can edit descriptive information about an image within the CMS. Some of this descriptive information is used to generate the ALT tag for the image.

4.4.4 File/page size

The eWebEditPro tool can only handle web pages of a certain size (in terms of HTML). Very long pages slow down the speed of the tool. However, this was considered to be an implicit advantage, as it would discourage CMS users from creating single web pages of a length that many people found difficult to load or read on screen.

4.4.5 File attachments

One of the more unusual aspects of the web site is that non-HTML files (such as PDF and RTF/ Word documents) are not linked to from the body of a document, but rather appear at the bottom of the document as a list of attachments. This was designed in response to complaints from many users that it could take a long time to "hunt the attachment" in long web pages, especially strategy and guidance documents.

A clear list of attachments at the foot of the page makes it easier to view and access all such files associated with a page. It also greatly reduces the amount of time it takes for disability browsers to get to an attachment, as a link at the top of the page enables users to jump straight to the list.

4.4.6 Printer version

Many users indicated a preference to read all but the shortest web pages offline and/or off screen. Therefore, a printer-friendly facility was developed. When activated, a version of the current page is opened in a new window with the navigation and template surrounding the content removed, enabling a "cleaner" version of the page to be printed.

4.5 Old web site content

A significant area of work involved moving content over from the old web site. This raised a myriad of issues; for example:

  • The lack of accurate metadata, or other descriptive information, for much of this content meant that new metadata had to be generated.
  • Decisions had to be made about old content that was written in the present tense e.g., "...for the latest version, contact..." where the version may now be out of print or irrelevant, or the contact person is no longer in post. The nature of some of this older content influenced the development of a number of archiving mechanisms (see section 5.5).
  • Some content did not easily fit into a category on the new web site. For example, some circulars appeared to a number of users to be news items, but to others the circulars appeared to be funding calls. In the new site, these older items had to be consistently categorized.
  • Inconsistencies in content location needed to be resolved. For example, in the old site it was discovered that there were several partial "indexes" of working groups. None of these lists alluded to the others, and therefore each gave the impression that they were the definitive / comprehensive list of JISC working groups. Users were therefore unaware of other working groups of potential interest. In the new web site, a single index of all current and previous working groups [9] was created so users could easily find all those of interest.

5. Finding relevant content quickly

In developing the new web site, the emphasis was on making it easier for people in the communities served by the JISC to find information of relevance. A wide range of considerations and opinion were taken into account, resulting in a variety of mechanisms being incorporated into the site, as described in the remainder of this section.

5.1 Main website sections

Links to eight main sections appear in the band header at the top of every page. These are sections or pages which are either in high demand or to which users required more visible links.

Various automated mechanisms were developed so that certain pages are automatically linked from a variety of locations. For example, when news items and events are entered into the CMS and published, then—so long as they are still "current"—a link to the relevant page is automatically generated and included on the home page from the title of the news item or event. When the item is no longer "current", e.g., the event has taken place, then the link and title disappear from the home page.

5.2 Popular pages

On the home page, a list of five links to pages on the web site is presented under the term "Popular". These are the five most downloaded pages over the last two weeks by all web site users. Superusers can edit a list of "Popular page exclusions" so that the home page, for example, is not always top of this list.

5.3 Quick links

Pages listed under "quick links" are those that have been downloaded most over the last two weeks from that particular PC. This feature requires the use of cookies to enable the system to identify the PC and generate the list of quick links from stored tabulations of accesses. A unique list will appear on each machine mapping the behaviour of the last user of the site. No personal information is collated using this method.

5.4 Guides to resources and services

Three different types of guide to Internet-based resources and JISC-funded services were incorporated into the web site: resource-oriented guides, user-oriented guides and topic-based "Advice and Guidance" information.

5.4.1 Resource guides

Resource guides [10] are developed from the perspective of specific academic subject areas. They offer an overview of the key resources available to the post-16 education sector. Each resource guide has been compiled by a subject specialist who, in consultation with members of his or her subject community, has selected, described and provided links to key resources in that discipline.

Seven resource guides have so far been developed, maintained by advisors based in various UK universities. The CMS was developed to allow the advisors to "share" a pool of common resources, and also to individualise the text to suit their own subject areas / guides. Mechanisms were built into the CMS allowing:

  • The collections team and the resource guide advisors to "share" common resources.
  • Trusted service providers to enter details of their resources into the CMS for consideration, editing and publishing by the appropriate resource guide advisors.

5.4.2 User guides

User guides [11] are developed from the perspective of specific types of user. These are web-only guides to services, publications, mailing lists, events, and other resources funded by the JISC. Initially, user guides were developed for those new to (or unfamiliar with) the JISC, technical staff, and staff based in FE institutions.

5.4.3 "Advice and Guidance" topic information

These [12] are developed from the perspective of "issues" or "situations" that academic staff may face. For example, one topic deals with issues surrounding the location, use and creation of electronic content. The topic sets out the key issues, and provides descriptions of mainly online resources and services that may assist in dealing with each issue. Topics can be downloaded as two-page PDF documents or browsed and searched in more detail on the web site.

5.5 Archiving mechanisms

A frequent complaint of web site users in the 2001 survey was the amount of "old" information on the web site that had to be negotiated in order to locate newer content. Three archiving mechanisms were incorporated into the system to make it easier to indicate to users content that is no longer being updated.

5.5.1 Automatic date archiving

In some sections of the website, when an item has exceeded a specified date entered by the CMS user, it is automatically relocated to a different section. For example, when an event has concluded, it is no longer linked from the current events [13] page, but is linked from the appropriate year in the past events [14] section. The same principle is applied to funding opportunities, publications, news and projects.

5.5.2 Manual archiving

With most files, the CMS user can "toggle" between the file being "current" (default) or "archived". Archived files appear on the live web site with a prominent "archived" banner [15]. (This banner links to the current JISC web retention policy.) In search results, manually archived items are also labeled as such.

5.5.3 Review archiving

Items can be assigned to one of a definable list of "types" within the CMS, e.g., briefing paper, consultation document or guidance document. Each type defines how long associated items are to remain archived before the web editor is notified by the CMS that the item should be reviewed, with a view to either further archiving or deletion.

5.6 Search mechanisms

Many users complained that the search system on the old site returned large numbers of irrelevant results. Consequently, a number of enhancements were made to the search mechanisms on the new web site:

  • Separate search systems were created for several areas of the web site, such as advice and guidance, collections, publications, and projects.
  • For speed of search results return and to avoid overwhelming the user, the site-wide search (available from every page) does not return every "hit" in one go, but in batches of roughly 8 pages of 10 results each.
  • Search results are ranked according to a percentage relevance score.
  • The indexing system indexes not just the web pages and their metadata, but also the content of several types of attachment, including Word and RTF files, PowerPoint slides and later versions of PDF files.
  • Pages set manually to archived status are clearly indicated in the search results. The user has the option of re-executing the search with archived pages removed. For example, a search across all pages on the word security produces 356 results; re-searching only across pages that are not archived returns only 33 results.
  • The user also has the option of re-executing the search with pages containing variations of search terms excluded.

5.7 Redirects

Some sections required redirects to take users to information in its new location. These have been used sparingly to avoid server overload and instead encourage:

  • Use of the site navigation and navigational aids
  • Utilization of the search engine and indexing system

Some content was reorganised into different sections as a result of JISC organisational restructuring. Redirects usually take users to a section "home" page linking to all the relevant content.

5.8 External search engines

At launch, the new site was registered with all the major search engines, such as Google. Procedures vary greatly for each engine registration, so for some the new site was not indexed for over 6 weeks. This did result in delivery of old site URLs during this time, but eventually only new site URLs were indexed. Therefore, external referencing greatly increased, as did user satisfaction.

6. Projects and Programmes

A substantial proportion of JISC funding and resources is committed to programmes of projects. Due to the large number of projects either completed or continuing, a separate section was built for programme and project information [16].

Programme managers were set up as CMS users, with the ability to control programme and project pages. Both types of pages have metadata fields within the CMS allowing the entry of specialist metadata, such as the programme/project start and end dates. Such information allows the web site to automatically generate browsable pages of current and completed programmes and projects, as well as forthcoming programmes.

As the JISC web site does not house individual project web sites (these are usually maintained at their host institution by the project team), a field is provided in the project description template for the URL of the project web site, which is displayed as a link from the project description web page.

The data structure for this section is built around the simple relationship between a project and a programme. A project is always associated with a programme (and conversely, a programme is associated with a number of projects). When a CMS user enters details of a project, he or she selects the programme with which the project is associated. This assists with project indexing; it also means that when a user views a programme description, the web page contains an automatically generated list of links to the projects funded under that programme [17]. Conversely, project description pages contain an automatically generated link to the associated programme description page.

User feedback indicated a variety of reasons for accessing information concerning JISC-funded projects. This necessitated the development of a variety of mechanisms to assist users in locating specific project information:

  • Browsable location: currently active projects can be browsed by programme, project start date, project end date, or by project title alphabetical listing.
  • Searchable location: searching across project descriptions, with optional filtering across a specific programme, and/or by whether the project is currently active or completed.

The web pages generated by all of these mechanisms use the description, metadata or other details stored about each project. For example, the options to list projects according to end date (useful when identifying projects that are soon to provide completed deliverables or final reports) uses the end date of projects that have yet to be complete.

7. Training CMS users

7.1 Installation and basic training

Most CMS training and support was undertaken remotely. Training consisted of:

  • Installation of the CMS on the PC used by the user, either by the user themselves or by an existing CMS user.
  • Downloading, by the CMS user, of the latest version of the CMS documentation. Substantial documentation was written in parallel with the development of the CMS and web site.
  • In some cases, introductory CMS training by the web development team.

CMS users took to the system in a variety of ways. More enthusiastic users with PC experience required virtually no training, and were editing and publishing content within 10 minutes of installing the CMS; other users required more detailed training and assistance. As JISC staff are housed mainly in four locations [18], JISC CMS users were encouraged to ask fellow users in the same office for help and advice. Previous CMS users were also useful sources of feedback regarding the functionality of the system.

7.2 CMS training facility

Within the CMS, an area was set up for inexperienced CMS users. This allowed people to experiment with creating, editing and publishing files, and creating CMS directories and navigational functions, without interfering with "public" content. The area is automatically wiped of files every night, therefore freeing up filenames for re-use and discouraging any CMS users from permanently storing content there.

7.3 Content quality

The tools and mechanisms available to the CMS user enable content to be processed and manipulated with considerable speed and ease. However, the CMS cannot improve the quality of such content. If a CMS user enters poorly written content, it will still appear on the public web site as poorly written content, albeit in a slightly prettier format.

A variety of quality issues have been raised, and new ones emerge periodically. For example:

  • Spelling control, despite the inclusion of a spell-checker in the CMS, is variable. Differences also exist in US and UK spellings of the same word, e.g., program and programme.
  • Staff have individual styles, for example regarding use of capital letters in terms such as Further Education.
  • Inappropriate use of formatting styles is sometimes found, e.g., using the same size heading for titles of different emphasis.

Evolving guidelines were provided in the documentation. However, it is clear that:

  • These guidelines will need to be periodically reviewed and updated.
  • Content produced by CMS users who are less quality-aware needs to be monitored, and extra training or support provided if required.

8. Future development

The CMS-based web site development was completed at the start of December 2002. Following formal approval from the JISC, it was made public in February 2003; in April, a full-time web editor was recruited to take over the maintenance of the web site and to provide non-technical support to CMS users.

Web site development is an ongoing, iterative process. There is still some content-based work to be done, such as collating and including project information from two early JISC programmes (NFF and eLib). Technical development work is ongoing, with a list of "desired" and "required" items kept constantly updated and prioritised.

Some additional functionality will shortly be applied to support CMS users, including:

  • Detailed, customisable web statistics
  • A link checking routine run as a service across the whole site
  • Implementation of a "web accelerator" to effectively hide the database string in the urls ('"index.cfm?name="'). This does not place a load on the server as redirects do, but allows more simple use of urls.

It is of paramount importance that the web site is developed according to the information needs of the community of users that the JISC serves. This ensures that:

  • The site is both useful for, and used by, those for whom it is intended.
  • Problems with the previous web site, such as the high prominence of irrelevant content, do not recur.
  • A high degree of content accessibility is maintained.

Feedback and comments about the JISC web site can be provided through a number of routes:

  • The enquiry and comment fill-in form [19]
  • Emailing the webmaster: <>
  • The appropriate JISC user group [20]
  • Comments made at JISC events and workshops [21]
  • Representation through appropriate support services, such as (for FE staff) the Regional Support Centres [22]


[1] The JISC web site <>.

[2] JISC web site development reports <>.

[3] The special educational needs and disability act, Sophie Corlett <>.

[4] Content Management Systems, Paul Browning and Mike Lowndes <>.

[5] Broadband <>.

[6] EduServ <>.

[7] Ektron eWebEditPro editing tools <>.

[8] Betsie (BBC Education Text to Speech Internet Enhancer) <>.

[9] Index of working groups <>.

[10] Resource guides <>.

[11] User guides <>.

[12] Advice and guidance topics and issues <>.

[13] Current JISC events <>.

[14] Past JISC events <>.

[15] Example of archived content <>.

[16] JISC projects and programmes <>.

[17] Example programme of projects: Authentication, Authorisation and Accounting <>.

[18] JISC offices <>.

[19] JISC web comment and enquiry form <>.

[20] JISC user groups <>.

[21] JISC events <>.

[22] Regional support centres <>.

Copyright © Clare Rogers and John Kirriemuir

Top | Contents
Search | Author Index | Title Index | Back Issues
Previous Article | Next Article
Home | E-mail the Editor


D-Lib Magazine Access Terms and Conditions

DOI: 10.1045/may2003-kirriemuir