Case Study: Staffordshire Past Track

Staffordshire Past Track is a web-based multimedia archive related to the history and cultural heritage of Staffordshire. The bulk of the collection is comprised of photographic or printed images from the past 150 years. Noteworthy features of the Past Track project are its bespoke content management system and a user interface that provides, among other functions, map-based searching, an image zoom facility, and user-annotated albums.

Prepared by JISC Digital Media with assistance from Andy Holt, Staffordshire Past Track Project Manager

Contents

1. Introduction
2. Project Outline
3. Creating the Content
4. Content Management and Delivery
5. Conclusions

1. Introduction

Like many digitisation projects, the Staffordshire Past Track project is not a simple project funded from one source and neatly time-bound. Although most of the work took place between 2000 and 2004, funded under the Lottery's New Opportunity Fund digitisation programme (NOF-digitise), Past Track built upon previous digitisation work and is likely to grow and be sustained through future projects and funding bids. This case study will concentrate on the NOF-funded work, but also seek to place it within the context of the wider programme.

The NOF work itself had two main parts: (a) the digitisation of content relating to Staffordshire; and (b) the development of a content management system (CMS) capable of storing and delivering the Staffordshire content and the content of several other West-Midlands NOF-funded projects. Each of these aspects is covered in some detail below (sections 3 and 4). The next section provides a context for the project and an overview of its history.

As this study describes a publicly accessible resource, you are strongly recommended to visit and familiarise yourself with the Staffordshire Past Track website before reading this case study.

2. Project Outline

The table below summarises the timetable for the Past Track project.

Project initiation, fundraising and early piloting phase (1999 - 2001)

In August 1999, the Heritage Lottery's New Opportunity Fund announced £50 million of funding towards the digitisation of UK cultural heritage resources in a programme called 'NOF-digitise'. (For a detailed background to this programme, see the article by Paul Miller and Sally Criddle in Cultivate Interactive 5).

Staffordshire Council already had some experience of digitisation: its Arts and Museum Service had previously digitised several thousand images for an interactive kiosk project. They were keen to gain NOF-digitise funding to build a more substantial, web-based resource. A partnership was formed to put together an initial bid. This was led by the council's Education Department, with membership drawn from several council departments and a number of other institutions and organisations from throughout the county.

In preparation for its application, the bid team examined the feasibility studies of other digitisation projects. The team also made a visit to the Hulton Getty image library in London to look at the way they had managed their digitisation and content delivery and to see how they were handling the digital rights management issues.

The council's initial 'outline' bid, submitted in early 2000, asked for £80,000 to build a 'Staffordshire Multimedia Archive' (see the summary of bid below).

Summary of the initial NOF-digitise bid

Main goal

• To digitise some of the disparate collections within Staffordshire, creating a substantial new electronic resource that reflects the rich cultural and community heritage of the county, provides enhanced access, and effectively exploits the potential offered by new information and communication technologies.

Specific Objectives

• Provide universal accessibility to regionally important learning resources
• Increase cultural awareness and identity
• Promote an understanding of community history
• Create a major sustainable resource
• Provide opportunities for community involvement

Audience

• Lifelong learners (a priority for NOF-digitise)
• Educational users
• Researchers
• Anyone with an interest in Staffordshire's past

Deliverables

• 20,000 digital images
• 200 video/audio files
• 20 themed web exhibit

While they were waiting for a decision from NOF, the council put together a prototype database, using staff and technical resources from its Education Joint IT Unit and two thousand images from the interactive kiosk project. This working prototype was regarded as the pilot or first phase of the Past Track project and was launched online in December 2000. It offered a browse and search, but lacked the sophisticated functionality that would be added during the NOF-funded phase (mapping, zooming and personalisation, described below). The main purpose of the prototype was to provide something tangible for its users to look at and comment on. It was by no means fixed and would change significantly over the next few years.

Staffordshire's bid was one of nearly 350 received by NOF-digitise, representing a total of £140 million. Given the high number of proposals, NOF chose to group similar bids together into consortia and ask them to submit revised bids based on shared resources. In early 2001 they grouped the Staffordshire bid with six other West Midlands applications to form the Sense of Place West Midlands consortium.

As five of the other consortium members required a content management system, Staffordshire Council offered to develop its prototype system into something the others could also use. This shared CMS formed a second part to the main consortium bid and was priced at nearly £170,000.

Work done during the pilot phase had suggested to the Staffordshire team (and to the other consortium partners) that their initial digitisation costings were unrealistic. As a result, the amount Staffordshire requested for its digitisation was more than doubled for the consortium Sense of Place bid: i.e. £164,000 for the same deliverables.

The main Sense of Place West Midlands consortium bid, which included Staffordshire's digitisation work, was accepted by NOF in mid-2001. The second part of the bid, the shared content management system, was not finally approved until September 2001. This made planning difficult and put more time pressure on the developers once work began.

Both parts of the project initially received less funding than requested, although there were further allocations as the project progressed. Although Staffordshire had asked for £164,000 for their digitisation work, NOF awarded them £80,000 in 2001. However, by the end of the NOF-digitise phase (January 2004), NOF had contributed a total £131,000 towards the digitisation work, to which the Staffordshire County Council had added a further £19,000. The final tally for the content management system was £183,000, all funded by NOF-digitise.

The NOF-digitise funded phase (late 2001 - early 2004)

Once the NOF funding had been approved, the programming team worked apace on the new version of the content management system. While this was proceeding, three new fulltime staff were recruited to form a digitisation team.

These two teams were divided between two sites and institutions. The CMS team were based within the Education Joint IT Unit in Stafford and the digitisation staff, at the County Museum at Shugborough Hall. Project management was shared between Andy Holt, an IT Manager within the Education Joint IT Unit, and Chris Copp, Museums Officer with Staffordshire Arts & Museums Service. Andy managed the development work in Stafford and provided line-management for the staff in both teams. Chris provided the Shugborough-based digitisation team with professional support and supervision. The PRINCE2 project management system was used as a framework for managing the project.

Staffing the Past Track Project

Project Management

Andy Holt, IT Manager (about 0.25 FTE)
Chris Copp, Museums Officer (about 0.1 FTE)

Digitisation team

Chris Copp (see above)
Howard Dixon, Digitisation Technician (Full Time)
Heather Dowler, Researcher/Data Coordinator (Full Time)
Alison Copeland, Researcher/Data Coordinator (Full Time) replaced by Jacqui Barker

Content Management System team

Andy Holt (see above)
Mark Gee, Senior Programmer (about 0.5 FTE 08/01-12/03)
Ian Dodds, Programmer (ad hoc)

The CMS developers were an established team who worked on this project over a concentrated period alongside other projects and responsibilities. The digitisation team were especially recruited for the project and worked on it full time from March 2002. Staff from each team were in regular contact, with full team meetings taking place each month to discuss progress and resolve any issues. These monthly meetings proved effective in addressing many of the issues related to project design and workflow. The larger issues the project faced included the following:

• Loss of staff member
  In mid-2003 the project lost one of its researcher/data coordinator staff members, who was, understandably, attracted by a permanent post available elsewhere. This can be a common problem faced by digitisation projects, which often have to hire staff on temporary contracts. As a result, the project incurred additional costs in recruitment and delays in the production as a new staff member was recruited and trained for the post.
• Repetitive nature of digitisation work
  The digitisation team became very bored towards the end of the NOF project, as a result of the highly repetitive nature of their work. As a result, members of the team sought and were given a more varied workload, which included some promotional work and educational and community outreach. While this lowered the number of resources being produced, it made for a happier working life.
• Delays in production of content
  The digitisation work was initially intended to run for 12 months, but faced delays - larger due to incomplete metadata and copyright information, and the difficulties of working with a large number of contributing organisations and individuals. NOF provided additional funding, which extended the digitisation phase to 21 months - until the beginning of 2004.

The development of the CMS proceeded with fewer hitches, but was complicated by the need to consult and agree the work with other consortium partners, some of whom had more limited experience with the technologies involved. In hindsight, the project developers felt they should have had more communication with the consortium partners earlier on in order to draw out their requirements and also to make it clear what was or wasn't feasible. For example, fairly late in the development phase, one of the partners asked if they could have a multilingual interface. While it might have been possible to build this into the project at an early stage, the requirement came too late for anything to be done without radically altering the structure and design of the content management system.

The project paid for some specialist consultancy, largely to confirm that they were on the right track and hadn't missed any requirements (and opportunities) from the standards and technologies they'd chosen. Metadata advice was provided by Stuart Holm (formerly of the mda and actively involved in the development of the UK Social History and Industrial Classification, SHIC) and Simon Tanner (then of the Higher Education Digitisation Service, HEDS). GIS/mapping consultancy was provided by ESRI (who supply the system's underlying mapping technology) and from Portsmouth University (specialists in historical mapping).

The team attended free NOF-digitise workshops and paid for hands-on training from JISC Digital Media. They also drew on several web resources, particularly:

• JISC Digital Media
• W3C
• NOF Digitise (no longer current)
• HEDS
• UKOLN

As the project timetable indicates (see above), the piloting of the web prototype continued into the NOF-funded phase. As the development work proceeded, new functions were added to the system and tried out on users. Workshops were held with a wide range of potential user groups at different locations and institutions throughout the county. These included local history clubs, teachers, school children, library users and the staff of libraries, archives and museums. Workshop participants were given an introduction to the website, allowed to explore it, and then asked to complete a questionnaire. More than 200 responses were gathered in this way and were influential in shaping the development of the system.

The project was formally launched in March 2003 although some of the development work and a large proportion of the digitised content was added after this date. As mentioned above, the NOF funding for the digitisation was extended until January 2004. Since February 2004 the digitisation team has continued to be funded by the Staffordshire County Council.

NOF-digitise deliverables

By the end of the NOF-funded project (January 2004), Staffordshire Past Track had publicly delivered:

• 10,000 digital images (7,000 photos and 3,000 art images)
• 75 video files and 3 audio files
• 10 virtual exhibits

This is less than the original estimates (see table above), especially as 5,000 of these images had been imported from previous projects. However, it is difficult to make direct comparisons, since the project did not receive all the funding it requested, and in January 2004 a further 2,500 images were being held back from the public while their metadata was finalised or copyright cleared.

Beyond NOF, sustaining the resource (2004 - )

Staffordshire County Council are continuing to fund the digitisation team because they value their work and the Past Track resource. However, the council's goal is to secure further external funding for digitisation. In addition, the project team are approaching other organisations to fund further development of the content management system.

Although there would seem to be some potential to sell the content management system to others, Staffordshire County Council haven't really tried to do so, citing issues relating to the funding of the original development and the need to ensure that all of the Sense of Place West Midlands consortium are considered. However, the team has tried to find other West Midlands organisations to join the existing partnership. Since the development work has been done and paid for, they are willing to set up a new partner for as little as £2000 plus a share of the running costs. These costs currently amount to £16,000 per annum for a 3-server platform, including hardware and software replacements and upgrades, and are split equally between the six consortium partners.

3. Creating the Content

This section summarises the workflow and the standards adopted for the project.

Some of the Past Track's contributing institutions refused to allow the material to leave their sites. In this case the Past Track team would do the work on site and, where possible, in 'one hit'. The researcher/data coordination staff would move in first, make the selection and begin entering data, working closely with museum staff. The technician would then come in with the scanning/photographic equipment, set up in an available space, and work through the material that had been selected. Depending on the amount of material, these visits might take several days or several weeks. Approximately 40% of the capture was done within partner institutions, with the remaining 60% brought in to the work space at Shugborough. While there are clear advantages to doing the digitisation work in a dedicated space, the visits provided some variety for the team and made the task of filling in the metadata gaps easier, since the relevant experts were on hand.

The project manager has, in hindsight, estimated that the project could have done with 1 scanning technician to 5 or 6 metadata and copyright officers. When the technician wasn't busy scanning he was taking digital photographs of objects, helping to code the static HTML pages on the site, providing general IT support and assisting with the transcription of metadata. In addition to filling in the slack time in image capture/processing, this helped provide some variety to his post.

Digital Capture

A proportion of Staffordshire Past Track images were inherited from previous projects. Some were of a lower standard than the Past Track project had set (i.e. they were of smaller pixel dimensions and saved in the lossey JPEG format). Some of the other imported images were at the other extreme, having been scanned full-size at 600dpi. The project reprocessed the imported images to integrate them with the new content as best as they could.

A small amount of large format scanning, such as maps, was outsourced to the University of Central England's Electronic Library service (UCEEL). UCEEL have Zuetschel large format book scanners, capable of capturing A1 (and above) at 800dpi.

The bulk of the Staffordshire Past Track content was captured during the NOF-digitise phase of the project using new equipment. Of these images, about 92% were captured using a flatbed scanner; 5% using a slide or transparency adaptor; 2% with a digital camera; and 1% were scanned and contributed digitally by others (e.g. website users). Like the imported images, these contributed images varied greatly in quality.

The digitisation team used the following equipment was used, chosen on the basis of independent reviews:

• Scanners
  Two flatbed scanners were used by the Past Track Project: the Epson Expression 1640xl, an A3 scanner with 1600x3200 samples per inch optical scanning resolution and 42-bit colour; and the professional version of the Epson Expression 1680, an A4 scanner with the same scanning resolution and 48-bit colour. The A4 scanner came with a transparency hood, which was used to capture negatives and slides.
• Digital Camera
  A Nikon D100 Digital SLR camera was also used for the project. This was capable of delivering 3,008 x 2,000 pixels, the equivalent of 6.1 megapixels. The camera was attached to a copy stand for large format overhead work and a tripod to capture works hung on walls. Three photographic lamps were used to light the works being captured. The camera was used to capture material that could not be scanned, due to its size, shape or fragility.

Scanners were calibrated and characterised using a Silverfast Ai IT8 Calibration kit, which had come with the scanners. This was done at approximately 6-monthly intervals and often resulted in the gamma settings being adjusted. ICC profiles were also created, which were used when the images were optimised. The white balance on the camera was checked and adjusted before capture.

The project used the following guidelines in capturing their images. Note that the outputs are expressed in terms of file sizes rather than scanning resolution or pixel dimensions (which might have been alternative approaches). These were guidelines rather than rigid rules: where the original held a lot of small detail, higher resolutions were used; where the originals were poor quality reproductions a lower resolution was sometimes chosen.

Media Type	Size of original	Bit depth*	Output Guide
Black and White Prints	6x4 inch	8-bit greyscale	4MB TIFF file
	8x6 inch	8-bit greyscale	5MB TIFF file
	A4	8-bit greyscale	7MB TIFF file
	A3	8-bit greyscale	10MB TIFF file
Colour Prints	6x4 inch	24-bit colour	12MB TIFF file
	8x6 inch	24-bit colour	15MB TIFF file
	A4	24-bit colour	21MB TIFF file
	A3	24-bit colour	30MB TIFF file
Glass Negatives (Black and White)	-	8-bit greyscale	10MB TIFF file
Black and White 35mm negative or slide	-	8-bit greyscale	5MB TIFF file
Colour 35mm negative or slide	-	24-bit colour	15MB TIFF file
Black and White Medium Format	-	8-bit greyscale	7MB TIFF file
Colour Medium Format	-	24-bit colour	21MB TIFF file
Note: *Where images contained fine detail within dark areas they were scanned at 12-bit grey or 36-bit colour and then downsampled to 8-bit/24-bit. Where a black or white original had developed colouration due to chemical changes, the image was scanned in greyscale. Where it had been deliberately tinted or coloured (e.g. Sepia), it was treated as a colour image and scanned in 24-bit colour.

Images were scanned into Photoshop via the SilverFast Scanning Application that came with the scanners. White and Black points were adjusted using prior to completing the final scan. Once in Photoshop, adjustments were kept to a minimum and generally limited to cropping. Occasionally rotation, deskewing, dust removal, brightness/contrast or levels adjustments were made. Optimisation was usually done immediately after scanning, so the original was available for comparison.

The raw images (directly from scanner or camera) were not kept. The optimised master images were saved to a server and backed up to magnetic tape as uncompressed TIFF v6.0 files. These files were write-protected to avoid them being overwritten when the surrogates (delivery images) were generated.

Digital surrogates, for screen delivery, were batch-generated in Photoshop some time after the scanning. Three images were created at this stage: a 110 pixel thumbnail image, a 500 pixel screen image, and a large 1500 pixel image which was used to generate a zoomable image file. These dimensions refer to the longest edge: height or width, depending on the orientation of the original work.

The only optimisation done to surrogate images was the application of Photosho's 'unsharp mask' filter which was used to compensate for the softening that naturally occurs when digital images are resized.

All of the surrogate images were batch saved as JPEGs using Photoshop. A quality setting of 40 was used for the thumbnails and screen images, with a better setting of 60 for the larger 1500 pixel image. Since this larger image was going to be delivered in a streamed zooming format, quality was more important and file size less crucial. Each item was given a unique number, which was included in its filename.

Someone other than the technician - usually the project manager or someone independent of the team - would check samples of the images to ensure their quality. This was generally done within a month of capture. A small number of images (less than 5%) required further correction or re-scanning.

Metadata Entry

The two 'Researcher/Data Coordinator' posts in the digitisation team took responsibility for the metadata and copyright clearance. They had professional experience in museums and archives and some experience with a previous digitisation project. As noted above, it took a lot longer to complete the metadata than was expected, mostly due to the incompleteness of the information accompanying the original images. The larger institutions involved tended to have card catalogue systems that were only beginning to be moved onto electronic systems. For many of the smaller collections, metadata consisted of little more than a hand-written or typed caption accompanying the image or artefact. Almost all of the metadata included in the final resource was re-keyed or created from scratch by the data coordinators.

The project developed its own metadata schema based on the Dublin Core standard. This schema was finalised through a series of workshops with project contributors and consortium partners. The schema was kept fairly simple and was focused on describing the original item (e.g. the photograph and, especially, the scene or content it depicts) rather than trying to make any complex distinction between the original work and the digital reproduction of that work, as some other metadata standards do.

In the absence of a suitable subject classification or thesaurus, the project devised its own hierarchy of subjects (these are the 'Themes' in the Interface). This was done in conjunction with two other members of the West Midlands consortium.

The metadata was stored in a relational database structure, with a main table for the 'Resource' and several linked tables which contained, for example, information about the donor of the item or the individual or organisation responsible for the digitisation work. The cataloguing interface screen shot (below) provides some sense of the metadata fields and tables.

In conforming to Dublin Core, Staffordshire Past Track has opened up the possibility of interoperating with other collections, through cross-searching or contributing records to other Dublin Core-compliant databases. With six projects sharing the same metadata and database structures, an obvious first stage of interoperability would be to enable a West Midlands cross-search. Although this is feasible, to date there has been no demand or interest in such a project. Nor has there yet been any demand from institutions with automated catalogues for links to be made with the Past Track. It seems likely, however, that cross-searching and catalogue-linking will be requested in the future.

4. Content Management and Delivery

Public Interface

The public interface is best experienced by looking at the site. The key functions to explore are the Theme Explorer, Map Explorer, advanced Text Search, User Album, and Zoom. The first four are accessed from tabs on the home page; the Zoom, from an individual item record.

The Theme Explorer enables users to browse through the collection by subject. Unlike many such browse interfaces, Past Track's retrieves the entire collection on the first click (it begins by displaying the first page, with a dozen items). Users can potentially browse through the entire collection page by page, but can also narrow down a topic of interest using a hierarchical tree, which operates like the Microsoft Windows Explorer.

The Map Explorer provides a clickable map with five levels of detail (from county to street). At any level the user can retrieve associated resources. Alternatively they can do a place or postcode search. The main mapping is Ordnance Survey data but the two most detailed levels offer additional historical mapping from 1775, c1900 and the late 1940s. Squares on the map grid are tinted in shades of yellow or orange to indicate the concentration of resources.

Level 2 of the Staffordshire Past Track's Map Explorer - click image for larger version:

Past Track's Text Search supports full Boolean searching (AND, OR, NEAR, NOT), phrase searching and "wildcards" (in this case, right-end truncation). Users can search on a date range or a broad historical period (when selected, these will automatically fill in the appropriate date range). Users are also able to limit the search to resources available for purchase (currently about 25% of the collection).

Once a resource is found (via the search, browse or map interfaces) users are able to view a screen-sized image and the associated descriptive metadata. From here they can 'place' the resource in their own personal album, adding their own annotations, or they can 'zoom' into the image to look at it in more detail. Users are also invited to correct any errors they see or supply further information about the resource or the scene it depicts. The project receives a couple of responses each day, which has led to richer metadata and the discovery of new collections within the community.

Behind the Scenes

The project's content management system is based on standard Microsoft software: Windows 2000 Server (operating system); SQL Server 2000 (Structured Query Language relational database); IIS (Internet Information Server, web server) and ASP (Active Server Pages scripting language). For security, Check Point's FireWall-1 is used. These technologies were chosen because of their widespread use and the developers' familiarity with them.

The mapping functionality was provided by ESRI's ArcIMS (Internet Map Server) product and the zooming functionality by Zoomify, both new technologies for the developers. These are discussed further below.

The CMS - and the content for Staffordshire Past Track and five other Sense of Place West Midlands projects - sits on three servers managed by Staffordshire's Education Joint IT Unit and connected to the Internet via the West Midlands Educational Broadband Network.

While the web-based system was being developed, the Past Track digitisation team (and other consortium projects) were given an offline application with a Visual Basic interface at the 'front end' and an Access database at the 'back end'. This had limited functionality, but served the purpose of enabling the projects to begin recording their metadata. Some consortium members became quite attached to this offline application and continued to use it once the online interface had been launched. However, with the addition of key functionality such as the mapping, zooming and thematic classification, the web version soon proved its worth.

The content management system supports three basic roles, each with different options and levels of permission:

• Administrators can create and manage collections (groups of resources), edit the themes (subject thesaurus), set permissions and access the user database.
• Editors can catalogue resources and manage several resource-related databases, such as the details of creators or donors, or the lists of resource formats
• Moderators can check and approve resources that have been created and edited by others

The screen shots below (click to enlarge) show the main cataloguing form and two of the additional cataloguing tools. The hierarchical theme tool is used to assign the resource to one or more subjects areas. This was originally built as a Java applet, but posed problems with firewalls so was re-written in Macromedia Flash.

First page of the cataloguing interface - click image for larger version:

Theme selection (Flash interface) - click image for larger version:

Map pop-up (based on ArcIMS software) - click image for larger version:

A map pop-up window enables the cataloguer to drill down through maps of various scales to 'pin-point' a location or area for the resource. Underlying the Map Explorer is GIS (Geographical Information System) software available from ESRI. The Ordnance Survey data was used within the terms of the County Council's licence at no additional expense to the project. Some of the historical mapping was licensed from Landmark and some was especially digitised and georectified for the project by the development team.

Like the metadata and copyright permission, the geo-referencing sometimes held up the delivery of resources, since a decision was made not to release an item until all the data and functionality had been included. Grid references were not available in any of the pre-existing metadata, so fresh coordinates had to be assigned to every resource.

Another technology the developers bought rather than made was the zooming functionality. The project users Zoomify, which is increasingly being seen in cultural heritage projects (e.g. the British Library, Getty, US National Gallery of Art). Zoomify takes in standard JPEG images, chops them up into tiles of different resolutions and then places them in a folder. A proprietary PFF file uses these to deliver a streamed, zoomable image, via a special viewer.

The extent of enlargement is generally 3 times (since the screen image is 500 pixels long and the zoomable image is generated from a 1500 pixel long image). However, where appropriate or useful larger images were sometimes included. The map shown in the example below could be enlarged by about 10 times (1000%) - enough to read all of its detail.

The full screen version of a resource - click image for larger version:

The same resource, enlarged with Zoomify - click image for larger version:

5. Conclusions

As this case study has suggested, large digitisation projects can prove challenging. They can be expensive and time consuming, subject to shifting technologies and standards, require many different types of expertise and many decisions.

Although the output of a digitisation project is standardised digital content, the inputs are far from uniform and can require a lot of work to be adequately captured and described. Although the technologies and much of the workflow are automated, many of the processes are not - nor are the human operators, who have their own needs and motivations. Digitisation projects become even more complex when they involve working with partners or collaborators, serve a broad user base, or, as in this project, additionally develop their own content management and delivery systems.

The Staffordshire Past track project encountered many of these issues. There were particular challenges in dealing with multiple partners (contributors and consortium members), in gathering metadata and securing copyright permission, and in motivating staff. These issues affected timetables and deliverables.

Given the project again, Andy Holt, the Past Track's project manager, said he would do much the same, but would put more effort into communication with partners and ensuring a variety of work for the digitisation staff.

He thought that time spent early on understanding the requirements of partners and discussing the limitations or consequences of these choices was time very well spent. He recommended having frequent face-to-face meetings and documenting them thoroughly, so you have something to fall back on. He felt that commercial partners could pose particular problems, since their salespeople were quick to say 'yes' to requirements and then worry about the practicalities later on.

It is essential, he said, that staff are given some variety in their work: "doing a digitisation project is not about setting up a production line, but about developing people as human beings." He felt that one way to provide some variety was to involve the digitisation staff in marketing the resource; useful since, in the course of their digitisation work, they will have acquired a great deal of knowledge about its content.

In a case study like this it is only possible to touch on a few aspects. If you would like more information about the Past Track project, please contact .(JavaScript must be enabled to view this email address).