On-line Collections Access at the Museum of English Rural Life

Jonathan Bowen, South Bank University, Roy Bridgen, Mary Dyson, Kevin Moran, The University of Reading, UK
Abstract

The Rural History Centre at the University of Reading, which includes the Museum of English Rural Life, holds extensive collections of material relating to the history of food, farming, and the countryside. As a result, it operates as the leading research and resource center for the subject in the country. The collections have been Designated of national importance and are now in receipt of a grant from the UK government-funded Designation Challenge Fund for a project designed to greatly enhance access to the collections via the World Wide Web. The aim is to have a dynamic, database-driven website which will present a thematic route of entry for exploration of the collections.

This collections access project aims to make the information available to a variety of users in an appropriate manner. Specifically targeted users are primary school children (and teachers), the general public (including secondary school children) and expert researchers (e.g., from academia). Different interfaces are provided within the website to accommodate these various types of user in the INTERnet Farm And Countryside Explorer (INTERFACE) section designated as schools, public and advanced users. For example, story-based environments are being included, extracting information from the database. The stories themselves are being generated by hand as static pages based on templates, but the database can be used at any time to include further information on objects or areas of interest. Primary school children can be limited to records that include graphical images, with access via story-based presentations suitable for use as educational resources, with associated teacher support. The public interface provides hierarchical exploration of subject areas. For experts, a more traditional form-based database interface with many fields will provide access to the full database.

One of the key considerations in developing on-line content for museums is bringing together collection management systems and exhibition information into one system. The manner in which collections are integrated into websites is likely to have implications for their accessibility. For example, the location of the access to collections within the overall site and the graphic presentation of the interface may influence visitors' search and navigation behavior. This paper includes information on an evaluation approach that was developed as part of the initial stages of the project. The objectives of the initial evaluation were to identify other sites which provide access to a collections database; to observe how functionality has been implemented; to consider which aspects of the interfaces may be adopted or adapted for the project. The outcome of the evaluations is described, identifying strengths and weaknesses of existing sites. As the sample of sites analyzed is small, this data is discussed in relation to potential uses for this methodology. A technical evaluation of the first version of the new website is also included.

The last part of the paper gives an overview of what is planned for the last part of the project. Not many museum collections are available on-line in a comprehensive form, especially from smaller museums, so this project aims to be exemplary of what can be achieved given appropriate resources

Introduction

The Rural History Centre (RHC) at the University of Reading is a national center for the study of the history of farming, food and the countryside. It began in 1951 as the Museum of English Rural Life (MERL) as a university museum (Boylan, 1999) when recognition of the impact of sweeping changes in agricultural practice led to the acquisition of material reflecting the pre-tractor era (Holt, 1977). Over the intervening half century, the collections have grown continuously and now include, in addition to the objects, a library of 50,000 books, one million photographic images, and extensive archival holdings relating to life and work in the countryside. The Rural History Centre today has many functions: as a university department engaged in teaching and research; as a public resource and information center on the subject; as a museum with a wide general audience, and special organized activities for school children.

In 1997, the Rural History Centre achieved "designation" status under a new government scheme to throw the spotlight on collections considered to be of national and international importance. It was followed by the creation of a Ł15 million ($22 million) Challenge Fund to be distributed amongst the 50 Designated Museums on improvement projects for the collections.

The response from the Rural History Centre was the Collections Access Project, designed to develop new forms of online access to its materials. This is a three-year scheme that began in the summer of 1999 with a total projected budget of just over Ł300,000 ($450,000). It fits within a broader picture of the Rural History Centre planning a move to new premises on the University site, reviewing its operations generally and recognizing the potential of the Internet for expanding access to the collections.

This paper presents an overview of this project. It includes more detailed information on evaluation studies that were undertaken at the start of the project and after the production of the initial website, an aspect of increasing concern to museums (Schweibenz, 2001). These looked at existing museum websites to help determine good and bad aspects of such sites and also used a questionnaire to obtain feedback from children using the website. More recently some technical evaluation of the initial website has been undertaken, especially with respect to enabling wider and easier access to the site. The project is on-going so the paper finishes with consideration of the project benefits so far and plans for the rest of the project.

The Project

The principal objective was to devise a means of online exploration of the collections, bearing in mind the diverse nature and requirements of potential users, that was capable of indefinite expansion and development. This immediately produced two main aspects to the project: the creation of an online database (Bowen, 2000c) for the collections and of a new front end operating over the top of it to deliver material dynamically in different forms for different audiences. Thereafter, both elements involved a number of contributory factors. The database already in use for the Centre's library had to be adapted to accommodate a whole range of non-book material from objects to photographs to archives and film. It is an Oracle based product called OLIB, licensed from Fretwell-Downing Informatics Ltd of Sheffield, UK. Once the cataloguing mechanics had been worked out, a plan was devised for the transfer of existing collection records on to the system. As the ability to view the collections online was important, a scheme for digitally photographing the collections was also put in train. Initially, an external photographer was contracted in to do the objects, while two-dimensional material was sent out to be scanned by a commercial firm. Increasingly, as the project has progressed, more of this work has been taken on by the in-house project team. High-resolution TIFF files are kept on CD-ROM, from which lower resolution JPEG files are created for mounting onto the server and linking to the database. The cataloguing and photography activities are ongoing through the duration of the project.

Creating a new front end to operate online over the top of the database required considerable thought and planning. From experience of the Centre's existing operations, three main categories of user were identified: (younger) school children; the general public (includes older school children); and specialists, whether they be academics, commercial researchers or members of the public with a particular interest. The new system needed to be able to serve up material from the database in a form suited to the needs of each category.

A simple subject heading system that would make navigation through the collections easily understood was devised. It moves from three primary headings - Farming, Country People and The Countryside - through secondary and tertiary lists of progressively more specific subjects. It was then possible to tag individual database records with their respective subject key words and with three separate and sequentially numbered captions, one for each user category. The system would then be able to deliver online a series of captioned images within each subject and user category.

For general and specialist users, the layout of the system is the same, operating on screen from the subject listings on the left hand side. They differ in that the caption sequence is longer in the specialist section and the amount of information given is greater. In the children's section, the approach is different. The starting point is a colorful picture of a fictional village and its surrounding countryside. When particular features, such as the blacksmith's shop or the farm, are "moused" over this activates the link to the relevant series of captioned images generated by the database. Additional interactive features, including a quiz, are also built into the village scene.

The great benefit of the system overall is its flexibility: the captioned image sequences can be extended indefinitely, the captions themselves can be altered at any time on the database and further subject headings can be added. In this way, it will continue to grow and develop along with the collections.

Packaging of the system online involved the complete re-design of the Rural History Centre's website in order to set it within a logical arrangement. A local Reading firm, Designation.com, was engaged to do this and to carry out all the associated design work on the new access system, which came to be called INTERFACE - standing for the Internet Farm and Countryside Explorer. From the new website [www.ruralhistory.org], INTERFACE is accessed directly from the home page. The raw data of the online catalogue can be viewed elsewhere from the Collections link on the site. Other links provide more information about the Centre and its activities; there is also provision for virtual exhibitions, the first of which will be developed during 2001.

The new website with the mark I version of INTERFACE went live in June 2000. Although the basic framework of the system is in place, much work remains to be done in filling out all the subject areas with a first round of captioned image sequences. A process of user evaluation has also been under way and an online feedback facility for comments has been incorporated. As a result, revisions and adjustments will be fed in at the end of 2000 in preparation for the launch of the Mark II INTERFACE in spring 2001. After that, the remaining twelve months of the project will be primarily taken up with further database input and sequence development.

The Project Team

The project has recruited a team of four young committed graduates, all in their twenties, who have been trained in the various component tasks - cataloguing, digitizing, authoring, evaluation, etc. - and have done a remarkable job in the short space of time to date. They are part of a wider team, which includes:

Members of the RHC's own specialist staff, especially the project manager, Dr Roy Brigden and the technician, Robin Harrison;

Two expert academic advisers, Dr Mary Dyson from the Department of Typography (Dyson et al., 2000) and Professor Jonathan Bowen of the School of Computing, Information Systems and Mathematics, South Bank University, London (Bowen, 1995-2000);

Pre-Development Evaluation Process

A preparatory stage of the project in the summer of 1999 provided a basis for future work by researching the existing situation with regard to web access to collections. This included a review of literature and an evaluation of a small sample of relevant sites. This research led to a set of recommendations and proposals on the information design of sites which provide access to collections databases, ways of dealing with multiple users, the generation of story-based environments, and how the collections database might be searched and returned (see Dyson and Moran, 2001).

In this paper, the framework developed to evaluate the usability and presentation of the sample of sites is described with summaries of example outcomes of the evaluation. The approach aims to provide a systematic method to identify and evaluate the dimensions that affect the usability of websites. The framework proposes criteria that may be considered in both formative and summative evaluations. These criteria may be translated into specific guidelines for the design of sites incorporating access to collections, but could also have more general application. The specific outcomes of the evaluations can provide direction in improving the design of websites by identifying good and bad practice.

The framework

A descriptive approach was adopted which combined the Systematic Usability Evaluation (SUE) of Garzotto, Matera and Paolini (1998) with aspects of a framework proposed by Walker, Reynolds and Edwards (1999). The resulting dimensions are modified to suit the purpose of the evaluation. This analysis is a variation on the method of heuristic evaluation [http://www.useit.com/papers/heuristic/] involving a systematic inspection of an interface. The heuristics are defined within the framework. This procedure is normally part of an iterative design process, and has been shown to be more successful with multiple evaluators. However, since the purpose of this evaluation was to identify strengths and weaknesses of existing sites to inform future designs, a single evaluator who was experienced in web design was considered appropriate.

A key feature of the SUE is that it considers the specific nature of the application to be evaluated, rather than addressing only general interface features. The method takes into account the structure of the information alongside the presentation and allows for discussion and interpretation of sites. Issues related to access and navigation, typography, use of images and text were taken from Walker et al. (1999).

Hampshire County Council Museums Service

Dimension	Strengths	Weaknesses
Consistency	Regular layout on each page. Consistent use of colour, type, type sizes.	Change in layout of pages when in 'search the collection' section.
Predictability	Results of search returned in similar manner regardless of class.
Accessibility		Difficult to get overview of site.
Typography	Headings clearly differentiated by using larger type sizes, colour, all caps or bold.	Line length can be too long.
Overall		Functional but not very engaging.

The Tate

Dimension	Strengths	Weaknesses
Consistency	Context of links within sections is consistent. Regular positioning of elements.	Screen needs to be slightly re-sized for some galleries.
Predictability	Collection ordering is coherent, i.e. order in list corresponds to logical order within site.	Preferable to use name of artist on links, rather than 'next artist'.
Accessibility	Site guide available.	Site guide can only be accessed from home page.
Orientation	Can navigate sequentially through lists of artists.
Typography	Line length is independent of browser window size.
Images & text	Good use of space and captions. Images can be decorative or informational and distinction is clear.	Cannot tell which images can be enlarged until on specific page.
Overall	Highly attractive and engaging through use of colour, space and structure.

National Gallery of Art, Washington DC

Dimension	Strengths	Weaknesses
Consistency	Body text has consistent typographic treatment.	Links to sections are sometimes before, sometimes after main text. Links have two graphic treatments: live text or gif images. Few graphic indications of separation between sections.
Predictability	Consistency between list and order of sections.
Accessibility	Index maintains the structure and hierarchy of the site. Plenty of clearly labelled links.	Option to search collection before searching site, which may confuse.
Orientation	Information is given on set of records currently shown in relation to total number of elements (e.g. 20-29 of 430).
Typography	Headings clearly marked. Typeface legible.	If browser window set too wide, line length too long.
Images & text	Information images are clearly labelled with dimensions, creators, dates etc.
Overall		Limited use of colour provides no aids to structure the site. Textured background can also affect legibility in some circumstances.

Evaluation of INTERFACE Mark 1

Once the initial website had been produced, evaluation of this was undertaken. The primary aim of this evaluation is to assess the effectiveness of the website interfaces to the collections in communicating to the intended audiences. In this context, effectiveness is defined in terms of:

At the time of preparing this paper, data has been collected from evaluating the school's section of the site. The online questionnaire to gain feedback from the general public has not yet produced many responses.

Method

A sample of 43 children between the ages of 8 and 13 years responded to questions about their general interests, preferences and experiences before answering questions about the RHC website. The children were either visiting the museum or were visited at their primary school.

Results

Two thirds of the sample use computers at home, with the remaining third using computers at school. These are mainly used with CD-ROMs (30%) or for accessing the Internet (28%).

The personal information collected from children on their interests etc. gives a good indication of the factors to be considered when planning the addition of new material to the site. The favorite subject at school is Art (27%), closely followed by Math (22%). This preference for Art might be exploited in the future design of new material. Quiz-like, logical problems may also be attractive.

A clear preference for novels (41%) over magazines (29%), comics (17%), factual books (10%) and newspapers (3%) supports the introduction of narrative elements to the web pages. The addition of facilities to create material, such as a diary entry, would also introduce greater interactivity. Although factual books appear to be unpopular, examples were cited as favorite books. Children may therefore not perceive such books as factual or may not understand the categorization. The style used in favorite books (e.g., the Horrible History series) may be an appropriate model for developing further material.

Many children had difficulty locating the children's section of the site from the home page. This will be remedied in the mark II version, by introducing a more obvious link. Once the INTERFACE page was reached, the instructions were considered easy to follow.

Technical Evaluation

As well as evaluation of the website from a user point of view, the initial website has been evaluated for its use across a range of browsers and to check its adherence to web standards. It is important that a museum website is user-centered (Bowen & Bowen, 2000) and also accessible by a wide range of browsers (Bowen, 2000b). A graphic design company who have moved into the area of web design designed the site. This means that the look-and-feel of the site is good. However, such companies tend to design for the currently most widely used web browsers, with little consideration given to lesser-used browsers. This is fine in an intranet situation, where most users may have a recent version of Internet Explorer for example. Even on the Internet itself, around 43% of clients visiting museum websites were using Internet Explorer version 5 in the year 2000 (Bowen, 2000b). Around 27% were using Netscape 4 and 20% Internet Explorer 4. This leaves about 10% of people using other browsers. Of course the situation is changing all the time. Netscape 6 is now available and access via WAP (Wireless Application Protocol) mobile phones may increase in the future.

Certainly the initial version of the website works well with the three most popular web browsers (see Figures 1 and 2 for the main home page and the initial INTERFACE page). Here we consider the use of the website with some other web browsers, its conformance to standards set by the World Wide Web Consortium (W3C) [www.w3.org], and its suitability for disabled access.

Browser access

A good browser for testing cross-compatibility is the Lynx text-based browser [lynx.browser.org] freely available on many operating system platforms, including Unix, DOS and Windows [www.fdisk.com/doslynx/lynxport.htm]. Accessing the main page of the current site results in the following display:

This is due to a "redirect" from the initial homepage to the "index.html" file containing the actual homepage, done automatically by most (but not all) web browsers. The http://www.ruralhistory.org/index.html link can be followed manually to obtain:

Here it becomes obvious that a large number of the images in the page have not been given "ALT" attribute tags to define alternative text (e.g., see "[shim.gif]", "[bw_2.gif]", etc., above). This is now a required attribute in recent Hypertext Markup Language (HTML) standards. Note that it is relatively easy to fix such problems. Here is an example of the same page, slightly modified to work more effectively with a wider range of web browsers:

Of course, the above page is not as attractive as the graphical presentation in Figure 1, but it is much more usable by those without advanced web browsers. It is also a considerable improvement over the previous Lynx screen.

Viewing the homepage with Amaya [www.w3.org/Amaya/], a web browser produce by W3C to promote the use of Extended Markup Language (XML), illustrates some further problems that may occur even with graphical browsers:

The alignment for the homepage is very critical since part of the graphic design (specifically, the pale colored rectangular potions of the display) is included as a "background" image to the page. This means that the rest of the graphic design must be laid out with margins of size zero at the top and left for correct alignment of the rest of the design. Both Internet Explorer and Netscape allow this using HTML attribute tags that have not been accepted in the latest HTML standards. Thus, it is highly likely that other browsers will misalign the design slightly, as illustrated above. Fortunately, this is not immediately obvious to many viewers, but the labeled square hyperlink images are positioned somewhat to the right and down a little compared with the paler background images.

Notice also that the font selection has not been recognized and the default Roman serif font is used instead at the bottom of the screen. It is important not to rely on a particular font typeface or size in web page design since these are not under the total control of the designer. Of course, selecting a preferred font, size, etc., is a good idea, and has been done in the design of this website.

Further potential problems include the use of "rollover" graphics using JavaScript where the graphical image changes as the mouse moves over a portion. Around 94% of museum web visitors have JavaScript enabled browsers (Bowen, 2000b), leaving 6% who do not. Rollovers provide an attractive addition to the web interface, but it is a good idea not to rely on their use for important aspects of any website. In Figure 2, the graphic at the top right of the screen is a navigation aid making use of rollovers. Fortunately, navigation is also possible via the text links at the bottom of the page, although it could be argued that these could be nearer the top of the page to allow those with small screens to access them without the need for scrolling to the bottom of the page.

Standards conformance

The standards for HTML coding are set by the World Wide Web Consortium (W3C). They provide an on-line HTML validator accessible under http://validator.w3.org/. The home page was tested using this validator. Below is a selection of errors discovered, and information on how these errors can be corrected:

It is quite common for the document type to be omitted from web pages. This is used to describe the type of markup used in the rest of the page. In a web context, HTML is assumed. However, it is helpful to include the DOCTYPE declaration at the start of web pages to give information on which version of HTML is being used in the page. For example, for version 4.0 of HTML, the following could be included:

There are a number of other possibilities, depending on whether strict or loose checking is required.

HTML is constantly in a state of transition. The latest versions of HTML (e.g., 4.01) are moving from using the "language" attribute tag to the "type" attribute tag for specifying the language used in any scripts included in a web page. For JavaScript, this following is now appropriate for backwards and forwards compatibility:

... f')" background="assets/home_back.gif" 
        leftmargin="0" topmargin="0" ma ...

... ack.gif" leftmargin="0" topmargin="0" 
        marginwidth="0" marginheight="0" ...

All of the above attribute tags have been introduced by Internet Explorer and/or Netscape, but have not been accepted by recent HTML standards. Thus although they work with the major browsers at the moment, this situation could easily change in the future. They have been included here to ensure the correct alignment of the graphical background and other graphical elements in the page design. A more reliable way to do this would be using Cascading Style Sheet (CCS) markup [www.w3.org/Style/CSS]. For example, the following could be included in the body markup:

This works with current widely used browsers and is more likely to work with future browsers.

The "ALT" attribute tag for alternative text with graphical images is now required by recent HTML standards. Many web pages, even those designed professionally, omit these. On the initial version of the home page, this error occurs no less than 52 times. Even if no text is appropriate, this should be explicitly included as "alt=""" since the default for a specific web browser is unknown and may not be the empty string (as for the Lynx browser for example where the file name of the image is included in square brackets by default). Further examples of this error are omitted for brevity.

All of the above errors are relatively easy to correct, once the problems are understood. Indeed, this has been done in a revised version of the homepage accessible on-line for experimentation purposes [www.museophile.sbu.ac.uk/merl]. Unfortunately, many tools used for the production of web pages do not adhere to official HTML standards, but rather conform to accepted practice by currently available web browsers at the time the tool was originally produced. The web is a fast changing environment, and using current HTML standards means there is more likelihood that the web page will continue to be rendered successfully with future browsers. Using ad hoc markup acceptable by current browsers gives no such assurance.

Disabled access

Museums generally pride themselves on disabled access to their facilities and in general try to avoid turning minorities away. The same has yet to translate itself in the context of the web to any large extent. Many museums attempt to produce websites that are accessible to the majority with standard browsers such as Internet Explorer and Netscape, without much thought for those with more unusual browser technologies. For example, audio web browsers aimed at the blind are now available (Bowen, 2000b). Unfortunately many web pages do not work well with these. The Bobby website [www.cast.org/bobby] provides a facility for checking accessibility of web pages. The main "Priority 1" accessibility error found with the existing home page is the lack of "ALT" attribute tags for alternative explanatory text with images. There are also five suggested user checks prompted by specific features in the page. It thus does not achieve "Bobby Approved" status as it stands. There are also three "Priority 2" accessibility errors and three "Priority 3" errors. Bobby also recognizes browser compatibility errors, as also found by the W3C validator. The TOPMARGIN, MARGINWIDTH, MARGINHEIGHT and LEFTMARGIN attributes in the BODY element are flagged as unknown in HTML4.0 and the ALT attribute is flagged as required for the IMG element (52 times).

Another useful indication provided by Bobby is an estimate of the download time for images, applets and other objects in the page and the HTML page itself on a 28.8 kilo-baud modem. This is summarized in the following table based on information taken directly from the Bobby results page:

URL	Size (Kbytes)	Time (secs)
http://www.ruralhistory.org/index.html	14.40	4.00
http://www.ruralhistory.org/assets/hed.gif	3.46	0.96
http://www.ruralhistory.org/assets/shim.gif	0.04	0.01
http://www.ruralhistory.org/assets/home_buttons/coll_pt1.gif	2.52	0.70
http://www.ruralhistory.org/assets/home_buttons/bw_2.gif	6.24	1.73
http://www.ruralhistory.org/assets/home_buttons/links_pt1.gif	1.90	0.53
http://www.ruralhistory.org/assets/home_buttons/edres_pt1.gif	3.35	0.93
http://www.ruralhistory.org/assets/home_buttons/centr_pt1.gif	2.27	0.63
http://www.ruralhistory.org/assets/home_buttons/intro_pt1.gif	2.89	0.80
http://www.ruralhistory.org/assets/home_buttons/intro_pt2.gif	4.29	1.19
http://www.ruralhistory.org/assets/home_buttons/bw_1.gif	5.46	1.52
http://www.ruralhistory.org/assets/home_buttons/intro_pt3.gif	5.00	1.39
http://www.ruralhistory.org/assets/home_buttons/intro_pt4.gif	3.92	1.09
http://www.ruralhistory.org/assets/reading_uni_shield.gif	0.94	0.26
http://www.ruralhistory.org/assets/made_at_designation.gif	0.89	0.25
Total	57.56	15.99
HTTP Request Delays	--	7.50
Total + Delays	--	23.49

The first column in the table above contains the URL of each item, the second column the size of the file in kilobytes, and the third column the approximate download time for each file on a 28.8 kilo-baud line. The homepage contains 14 images (some used more than once, but local caching means these do not need to be reloaded). An arbitrary 0.5-second delay is added for each file to allow for Hypertext Transfer Protocol (HTTP) connection times. In this case, the total delay is around 23.5 seconds for 58 kilobytes of material. A more modern 56 kilo-baud line would halve this delay to about 12 seconds, assuming no delays are introduced elsewhere on the Internet. This is on the high side, but is reasonably acceptable.

This technical evaluation has revealed a large number of problems, but most are fortunately minor and easily correctable. None of the problems encountered during the various evaluations is insoluble and the project has the opportunity to improve access to the resources in the final stages of the project. Evaluation is certainly a very important aspect of the production of a professional website that should not be overlooked by museums.

Conclusion and Future Plans

The project continues to run until 2002, but already a range of benefits can be identified. Above all, the project has facilitated a major re-think on the development of the Rural History Centre's services for the future. Extending access to collections is currently a key political maxim in the UK museums sector and a vital key for unlocking sources of funding. The Internet is seen as having great potential in this respect and there are a number of government inspired initiatives under way designed to make more cultural and educational material available online through, amongst other things, a National Grid for Learning. By means of its Collections Access Project, the Rural History Centre has been able to move directly into the mainstream of these developments. The necessary pre-requisites, namely the wholesale transfer of manually held records onto database and the production of digital image files, could hardly have been attempted without the special funding that the project brought. The launch of a new website putting the Rural History Centre online as a hub for information, resources and links on the history of food, farming and the countryside also has beneficial spin-offs for the emerging network of rural and agricultural museums in the UK. The co-operative elements within the project have also been very important. The whole enterprise has been an exercise, or an experiment, in working together across a range of disciplines and specialisms, producing a number of papers and reports (e.g., Dyson et al., 2000) along the way for future reference.

The key objective for the final year is to make full use of the expertise now built into the in-house team to consolidate and complete this phase of the project. Some of the principal tasks envisaged include:

Online cataloguing. This will be the role of the Cataloguer and Assistant Cataloguer. Both are fully trained, achieving the target outputs of 65 objects or 100 photographs each week. This material goes on to our OLIB database, the online version of which is available from the RHC website.

Digital photography. The Project Assistant will continue to create digital images of objects in the collection, to feed into the system. The target will be a further 2,500.

Scanning images from the photographic and archive collections. Some of this will be done externally, the rest in-house on a piecework basis. It is hoped that at least 10,000 new image files will be added.

Population of the existing subject areas within the INTERFACE system. Currently there are a total of 51 subject headings on the system, each of which by the end of the project will have five database-generated captioned image entries at the General Public level and ten at the Specialist level. The production of these caption sequences is the single most time-consuming element in the whole project. It will therefore be an important feature of the final year, occupying much of the available time of two of the project team and the project manager.

Evaluation. This has been built into the project from the beginning and feedback from it has informed the changes that are to be implemented in the INTERFACE system. We shall be carrying out some further evaluation, gathering and collating comments and returns made by users. There will also be an overall evaluation of the project itself. The whole scheme has been a major undertaking for the Rural History Centre and there will be lessons to be usefully learned in subsequent work.

The web continues to develop apace (Berners-Lee, 1999) and museums with continue to be affected by this (Bowen, 2000b; Farnell, 2000). Future interesting possibilities for the Rural History Centre website include learning and collaboration (Bowen & Houghton, 1999), visual search as opposed to text search for photographs and other images (Lew, 2000) and e-commerce activities (Bowen, 2000b). The need for efficient provision of ever-growing image resources on-line will increase (Talagala et al., 2000). The end of the current project will by no means be the end of on-line developments for the Centre, which must always aim to keep abreast, and where finances permit at the forefront, of the ever-changing world of the web.

References

Bowen, J.P. (2000c). Collections databases. New Heritage, 04.00:34, July 2000.

Dyson, M.C., Brigden, R., Bowen, J.P., Jenkins, K., Palmer M. and Phillips, W. (2000). Interfacing Disciplines in the Design of Web Interfaces. In S.A.R. Scrivener, L.J. Ball and A. Woodcock, CoDesigning 2000: Adjunction Proceedings. Conference of the Design Research Society with the Design Education Association (DEED), Coventry, UK, 11-13 September 2000. University of Coventry, pages 35-40.

Dyson, M.C. and Moran K. (2001). Informing the design of web interfaces to museum collections. Museum Management and Curatorship. To appear.

Farnell, G., editor (2000). Heritage on-line: Special report. New Heritage, 06.00:17-36, November 2000.

Holt, J.C. (1977). The University of Reading: The first fifty years. Reading: Reading University Press.

Lew, M.S. (2000). Next-generation web searches for visual content. IEEE Computer 33(11):46-53, November 2000.

Talagala, N., Asami, S., Patterson, D., Futernick, B. and Hart, D. (2000). The art of massive storage: A web image archive. IEEE Computer 33(11):46-53, November 2000.

Walker, S., Reynolds, L. and Edwards, V. (1999). Interactive multimedia in primary schools: children's use and understanding of information texts on CD-ROM,