Welcome to peterboyd.com
Peter D. A. Boyd
'Darwin Country' - a case study in website creation and content management
Peter D. A. Boyd
Collections Manager and Darwin Country Project Coordinator, Shrewsbury Museums Service
BOYD, P.D.A. (Forthcoming) 'Darwin Country' - a case study in website creation and content management. In Special Content Management Volume of Spectra ( a publication of the Museum Computer Network)
'Darwin Country' is a British website project initiated in February 2000. It was grant-aided by the Resource IT Challenge Fund and the West Midlands Regional Museum Council until March 2001 and, with funding for additional digitisation, by WMRMC April 2001- March 2002.
This database-driven website provides multi-disciplinary content in what the author has called a 'knowledge-net' environment. The content is stored in a mySQL database running on a LINUX machine with Apache as the web server. Pages are created using queries on the database server using PHP.
The first stage of the project was called 'Cradle of Science, Technology and the Better Life!' It was intended to provide an introduction to scientific, technological and social development in part of the West Midlands of England during the 18th and 19th centuries illustrated by the archives, paintings, decorative arts, archaeology and scientific collections of the partner museums. The initial partners were Shrewsbury Museums Service (the lead partner), Ironbridge Gorge Museum Trust and the Wedgwood Museum.
The scope of Darwin Country was extended after March 2001 to include material from other periods and by April 2002 the website had about 9000 pages and 6000 images. This paper provides an update on development of the website as described in Boyd (2002) and further information on the concept and mechanisms of Content Management implemented during the course of the project.
In 1999, the DCMS/MGC IT Challenge Fund invited applications from museums for the creation of innovative ICT projects that would provide content for Lifelong Learners based on material in museum collections. In September 1999, the author put together an application for Shrewsbury Museums Service to lead a project named 'Cradle of Science, Technology and the Better Life' supported by two museum partners, Ironbridge Gorge Museum Trust and the Wedgwood Museum. The project was one of 11 projects to be successful in attracting grant-aid and work started in February 2000.
Note: In April 2000, the Museums and Galleries Commission (MGC) became part of the new cross-domain quasi- government body for museums, libraries and archives, 'Resource', and the fund became known as the 'DCMS/Resource IT Challenge Fund' - shortened in normal use to 'Resource IT Challenge Fund'.
Key aims of the project (which became the first stage of Darwin Country)
Key objective of the project
In part because the 19th century naturalist Charles Darwin was born and raised in Shrewsbury, the project was subsequently named 'Darwin Country' with the website at http://www.darwincountry.org/.
Associated aims and objectives of the project
The Museum Partners:
Website Content Management is often discussed in the narrow terms of a software system for managing the content of a website but, in reality, it is part of a process that may include (among others):-
· a Vision (a fairly clear idea of what you want to do and how you want to do it)
· a Strategy to put that vision into practice (e.g. who does what where and how?)
· a Software Solution that incorporates a Content Management System (CMS)
· Content Creation by photographers and authors
· Upload of content to the website via the CMS and/or FTP
· Coordination by a Manager to make sure that the different elements work together to make the vision a reality
The detailed aims and objectives of the initial project have been given above but the following concepts embody important basic elements of the author’s vision for the initial and continuing project:-
· make museum objects accessible that are rarely seen by the general public because of the limitations of physical space, sensitivity of objects to light or their condition
· show details of objects not normally visible in a museum display (because they are on the side away from the observer or the details are too small or too distant to be seen)
· draw attention to details of an item to delight the observer and/or improve its interpretation
· 'display' items in a context or contexts that add to interpretation and enrich experience or understanding
· make associations between different objects (even from different museums) that would be difficult in a physical space and associations with the physical world outside the museum
· make the experience of accessing the website a multi-disciplinary one with opportunities to discover links between the natural and man-made world
· allow people to access images of and information about museum objects even though they may be unable to visit a museum in person because of disability or distance from it
Development of the Vision
From this it is will be seen that we were keen to develop a website that did more than just make information about a large number of objects available. While this alone would play an important part in improving access to our collections, the web provides a way of doing things that could not be done through a museum display or a book. We wanted to make it possible for users of the website to explore its contents in several different ways including interrogation through a search engine, interactive maps and interdisciplinary themes.
We wanted to provide gateways to enable visitors to find out about people, places, industries, ideas and other topics of the period but there were a number of interdisciplinary themes that we wished to take into account in the development of the website including:-
However, we decided that if we were to create a website structured according to themes that were defined too precisely, those themes would limit the future development of the website. A solution was sought that would allow an indefinite number of themes to be developed.
Key elements in the strategy and management of the Project
In the initial stages of the project, we had been thinking in terms of the site being made up of 'hand-crafted' web pages. However, it became apparent that only a database-driven solution would enable us to develop a website that was flexible, editable and updateable in the way we wanted. It was also apparent that no existing museum websites had the capacity to allow for multidisciplinary content and connectivity in the way envisaged and no parallel was found in websites of other disciplines.
Website design companies were investigated by conducting searches on the web. An initial assessment of what each company did was made through their websites and the way in which they responded to enquiries. We selected Orangeleaf Systems Ltd, a Shrewsbury company of software engineers, to work with us in developing an 'Internet Solution'. Orangeleaf Systems specialised in database-driven websites and writing software to control complex manufacturing processes and one member of the company had experience working with museum IT projects. We were remarkably fortunate to find that the most suitable company had recently moved to Shrewsbury and it was very helpful to be able to discuss matters face to face whenever necessary - and collect or deliver software or data discs within minutes.
The website 'solution' sought was based on the desire to provide access to images of museum objects in a multi-disciplinary way and create what I have called a 'knowledge-net'. Initially, discussion of 'the solution' that would allow this was focussed on database structures created by the author in MS Access. This helped to provide some common basis for discussion because, at first, the author as a museum person and the software engineers were speaking a different language and using the same words in different ways (e.g. a website object is not the same as a museum object). There was a steep learning curve for the author to look at things from a different perspective if a database was to work as a web page generator and for Orangeleaf learning about museum and multi-disciplinary perspectives.
It was necessary for both sides to look at things in new ways and work towards a common use of language as well as 'the solution'. It was a stimulating experience for both sides. The willingness of Orangeleaf to write software to do things that they had not been asked to do before or that, as far as they were aware, nobody had been asked to do before was key! As a result, after about two months work, Orangeleaf Systems succeeded in providing the first version of a website structure and Content Management System to make the author’s vision work.
We decided that the solution should conform to the technical standards identified as Best Practice for the nof-digitise part of the New Opportunities Fund (a British grant-aid scheme for providing website content from museums, libraries and archives launched after the IT Challenge Fund). In addition, we created our own standards regards image sizes and added features for usability such as the ability to increase the text size on a page with a single mouse click for those with visual impairment.
The pages of Darwin Country are generated dynamically. The content of the site is stored in a mySQL database running on a LINUX machine with Apache as the web server. The pages are created using queries on the database server using the open source server side language PHP, using just under 20,000 lines of code. Some graphical interface work was performed using Perl 5™.
To achieve the what the author has called the ‘knowledge net’ structure, the main content of the website is organised as 'Categories', 'Objects', 'Events' and 'Images'.
The top categories were chosen with care to allow anything within Shrewsbury Museums or others to be capable of being accommodated. They are:-
These are the eight points on the compass rose on the Home Page (Fig 1) and they appear at the top of each page on subsequent pages of the website (Fig 2). There is also a Search facility based on words within the name of a category, object or event or keywords that have been associated with the category, object or event.
The names of places such as 'Shrewsbury', people such as 'Charles Darwin', plants such as 'Passion Flower', animals such as 'Butterflies', objects such as 'Coalport Porcelain' or topics such as 'Pteridomania' are all 'categories' with their own pages and related or associated objects.
By April 2002, there were about 3000 different categories and 6000 objects (each with three sizes of images). This equates to about 9000 pages on the website. Each category can be associated with up to seven other categories and each object or event can be associated with up to seven categories.
In this way a 'knowledge-net' structure (Fig 4) is created, providing access to information via a large number of different routes rather than the single defined route of a hierarchical-structured website (Fig 3).
A category such as 'Pteridomania' can have up to three images within the text of the page and associated objects provide thumbnails of associated museum objects (with brief description) that relate to the page - these are hundreds of objects in some cases (fig 5). Clicking on the thumbnail opens a new page with a larger image of the object and the full description. Clicking on this image opens a new page with an even larger image. Therefore 6000 different images equate to 18,000 image files on the website (thumbnail, medium and large).
Events provide the data for timelines that are generated on pages that have been actively associated with those pages (e.g. Darwin events associated with Darwin pages) or automatically through a start and end time when events are associated with a linked lifetime of a person or period (fig 6).
Interactive maps and web-based GIS
Shrewsbury Museums Service already used a Geographical Information System (MapInfo Professional) in association with its MS Access-based database known as ShrewD (Boyd, 1999). Web-based GIS was a development of this. Interactive map-based web pages were intended to be one of the ways into information on the website. Web-based GIS can be a much more expensive technology than desktop GIS but it need not be. The author has discussed some of the key issues for GIS on heritage websites elsewhere (Boyd, 2000).
Copyright licensing costs on map data can be prohibitive in the UK. However, digitised map data from Europa Technologies was purchased at a special rate with an agreement to allow us to use a certain number of maps (with our own superimposed data) on the web without charge. Interactive maps have been generated using this map data and MapInfo Professional™ with web.Publisher™ linking to pages of Darwin Country already created. In this way, the interactive maps provide access to pages (Categories or Objects) that are opened when the user clicks upon hotspots on the map (fig 7). It is the intention to further develop this feature with more detailed maps that have been digitised for the purpose.
During the course of the project, the UK Ordnance Survey has slightly relaxed the use of their map data on the web for Local Government users but their rules are still more restrictive than one would wish for educational users.
Practical aspects of website content creation and project management
When we had achieved the mechanism to deliver information via the web we had to start generating that information and getting it onto the website. While some information could be gathered beforehand, you have to have to know the mechanism you will use for delivery before you know how to create and organise the information.
Practical content creation processes include:-
· selection of museum objects
· photography/digitisation of museum objects (whether 2D or 3D)
· uploading images (thumbnails, medium and large) to the website
· preparing primary object image record on the website database using Content Management System so that image is displayed on the site
· enhancing object image record with text, keywords and category links - may need to create new categories
· making object records live
· monitoring development of the website so that additional cross-links may be made between categories and objects created at an early stage as new categories are created at a later date
The Content Management System allows a public view of the website with those pages and images that have been made live and a non-public Administration view with both 'live' (public) pages and those that have not yet been made 'live'. Creation and editing of records (pages) can be done before pages are made public but we have continued with a procedure that was implemented during the initial stages of the project with the encouragement of the IT Challenge Fund - that of making new images public as soon as possible with minimal information and gradually enhancing that information 'on the fly'.
Selection of museum objects
In the first stage project, the objects to be photographed were selected by each partner museum according to the primary aims, objectives and identified 'themes'. The scope of the project has since been enlarged to include all periods and types of museum object. The selection of objects has been based on working towards the 'completion' of certain key collections (e.g. Shrewsbury Museums collection of paintings), providing access to images of every item in certain 'new' types of discrete collection that are not readily accessible to the public (e.g. Shrewsbury Museums collections of Trade Tokens and Commemorative Medallions), increasing the diversity of objects depicted especially those not readily accessible to the public (e.g. herbarium sheets, microscope slides) and starting the digitisation of important large collections (e.g. fossils, archaeological items from the site of the Roman town of Viroconium at Wroxeter, near Shrewsbury) that were not included in the scope of the original project. Some of these elements of the museum collections have never been properly documented before so that this is seen as an additional merit in tackling those collections.
Photography and digitisation of museum items
Accessing and preparing large numbers of items for digitisation and ensuring that the associated documentation is accurate (to ensure a smooth transfer of information to the website database) may take longer than expected.
As the professional photographer employed to photograph the museum objects is paid by the day, it is essential to manage the process effectively to get the best value from it. We have found that it is necessary to have enough objects ready to allow the capture of approximately 200 images in the day. If the objects are all of a similar size and easy to put down and photograph, the number can rise to about 240 per day but if the objects vary in size and shape and create different reflections (e.g. silver objects) requiring frequent adjustment of camera position and lights, the number achieved may fall to 160. Some objects require only one image but others require several to capture different details. You have to have in your mind what you are going to want to draw-out on the website – the more efficient you are on the day, the more that will be achieved for your money and there will be less need to re-insert an object for additional photography on another day’s shooting.
At least, digital photography allows you to see what you have got as you go - so you do not have the uncertainty of traditional photography in which you may find too late that the exposure was not right or a disaster happened in the darkroom! It is also much cheaper than traditional photography – the cost of film or plates and processing of 200 images each session would be prohibitive for us. The initial material cost of 200 digital images is nil!
An independent professional photographer, David Houlston, has done most of the digitisation for the project using the most up-to-date digital cameras and flatbed scanners. As a professional photographer, he has to keep his equipment up-to-date in a way that we could not. The digitisation has been done at high resolution (about 20MB) and he provides high-resolution images for archiving with images optimised at three different sizes/resolutions for use on the website. His up-to-date equipment and skill ensure that we get more consistent high-quality results at a lower price than we could achieve in any other way. Images have been optimised and otherwise manipulated using mainly Adobe Photoshop™.
We settled on the following recommended maximum sizes for most images:
Thumbnails 5Kb (150 x 150 pixels)
Medium 20Kb (500 x 500 pixels)
Large 50Kb (2000 x 2000 pixels)
However, in practice, a few images were found to contain so much detail that the medium and large images had to be larger file sizes to display well without loss in image quality.
With the first block of images to be captured, we based the image file name on the museum object accession number. This was a mistake and it led to unwieldy, unfriendly and confusing file names, particularly when one object might have several different images associated with it. All subsequent images were given their own simple unique number (e.g. sy1264, ig342) with the prefix identifying the museum from which the object came.
Hundreds of images may be uploaded to the website at one time using CuteFTP™ or one at a time using the Content Management System provided by Orangeleaf Systems. In cases where large numbers of images have been made ready, the optimised images are uploaded to the website using FTP before the object records that relate to them have been created using the Content Management System.
Creating records using the Content Management System
The Home Page of the administrative non-public part of Darwin Country is shown in Fig 8.
Records are made up of:-
The Admin Home Page provides access to the database tables for creation and management of these as well as other features for managing the website.
Fig 9 shows part of the information input form for 'Categories' that is the way in which 'Category' pages such as 'Pteridomania' are defined. Fig 10 shows part of an information input form for 'Objects' that is the way in which individual images are defined. These 'objects' may be one of several images relating to a single museum object. An 'Event' input form is similar.
Each 'category', 'object' or 'event' may be linked to up to 7 categories using the form. This provides part of the basic mechanism for creating the knowledge net structure of the site. In this way, thumbnails and information about each image may appear as 'Associated Objects' on up to 7 different pages (categories).
Therefore, for example, the image of a painting may appear on pages that relate to:-
· Fine Art
· the artist who painted it;
· the medium (e.g. watercolours)
· period (e.g. Victorian Period)
· the subject (e.g. Shrewsbury Abbey)
· type of subject (e.g. Abbeys)
· the place (Shrewsbury).
Sometimes, 7 categories are not enough and a decision has to be made about which associations are the most important. When the website was set up it was decided that 7 gave enough possibilities for the system without the SQL engine having ‘heart failure’ every time that it was asked to create a page. When each page is opened on the website a very large number of queries has to be processed to be able to display the page with its associated objects and categories.
Monitoring development of the website
As the website is developed, new categories are created or those created earlier made ‘live’. Therefore, it is necessary to revisit ‘records’ made earlier in the project to create new and/or different links. Categories, Objects or Events can be re-worded without links being broken because the unique ID number of each ‘record’ is what is used by the system – not the words!
Also any record may be enhanced by additional keywords, revised text or different images at any time. This is the joy of a Content Management System – if you make a mistake you can correct it - if you have a new idea you can incorporate it!
Website preservation strategy
Shrewsbury Museums Service has a commitment to Darwin Country and its future development. We intend that the site will continue to be a living, growing thing. It is therefore important to ensure that the website can be resurrected if the server blows up or if some other minor or more major disaster places it at risk.
As well as copies of the data of the website being saved automatically from the server every 15 minutes, copies are also downloaded ‘manually’ at intervals onto PCs in different buildings and different dated versions retained. A full copy of the Darwin Country website including all software and data including images is kept on a laptop computer and another copy is available as a touch-screen version for in-gallery use. The off-line versions are updated at intervals with the data and images.
In addition to the above, all the images are held on desktop PCs and laptop, on CD and tape-streamer with copies in different buildings. Images will be migrated to new storage media as it becomes advisable.
Usage of the site
Use of the website is monitored using a statistics package set up by Orangeleaf Systems as part of the Content Management System. This records number of page impressions, number of unique visitors and other information such as “most visited pages.” The package does not measure hits to the administration part of the site that is used in inputting data, so the page impressions logged represent other users.
In the 12-month period April 2001-March 2002, the website achieved about 60,000 page impressions (after editing out of ‘hits’ by over-zealous search-engine spiders) and about 10,000 unique visitors from all over the world. This is considered to be a pleasing start but, as information on usage of other museum websites has not been readily available, it has been difficult to compare our performance with other sites.
The top 20 countries accessing Darwin Country in the period January - April 2002 were the United Kingdom, United States, Netherlands, Canada, Australia, Denmark, France, Japan, Italy, Spain, Sweden, Belgium, Greece, Germany, Poland, New Zealand, Brazil, Austria, Singapore, and Norway.
The Darwin Country project has created a website that already has considerable value as a resource for life-long learning with 9000 pages and 6000 images available.
Its Content Management System has allowed:-
· Flexibility – change things and add new elements – on the fly!
· Manageability - growth/development possible without specialist IT staff
· Sustainability - we can develop the website at as slow or fast a rate as our resources allow
Boyd, P.D.A. 2002. 'Darwin Country - Cradle of Science, Technology and the Better Life!' - a case study In mda information vol 5, no5, pp 35-44. Proceedings of the Museum Documentation Association Conference, September 2000: ‘Clicks and Mortar - building cultural spaces for the 21st century’. [Available at: http://www.peterboyd.com/mda2000.htm]
Boyd, P.D.A. 2000. Heritage UK plc - a portal to space and time through web GIS. In Geographic Information supporting UK plc. Association for Geographic Information Conference Proceedings 2000. Published on CD. [Available at: http://peterboyd.com/heritageUKplc.htm]
Boyd, P.D.A. 1999. GIS in Museums - a case study. In Access to Better Information. Association for Geographic Information Conference Proceedings 1999, 8.9.1-8.9.7. [Available at: http://www.peterboyd.com/AGI99.htm]
Darwin Country website at http://www.darwincountry.org
Shrewsbury Museums Service Website at http://www.shrewsburymuseums.com
Orangeleaf Systems website at http://www.orangeleaf.com
Peter D. A. Boyd's website at http://www.peterboyd.com
National Grid for Learning at http://www.ngfl.gov.uk/index.jsp
Nof-digitise Projects Technical Information Papers at http://www.ukoln.ac.uk/nof/support/help/papers/
Figures (web versions not available at present)
Fig 1. Screen-shot of database-driven Home Page of Darwin Country website
Fig 2. Screen-shot of part of database-driven page of Darwin Country showing a top level category 'Places' page linking to other place-related ‘categories’.
Fig 3. Conceptual model of 'hierarchical' database-driven website showing 'hierarchical' structure of most websites whether 'hand-crafted' or database-driven. Each circle represents a page or link so that navigating a website from a link on the Home Page (at the top), the routes are pre-defined by the designer of the website. The visitor can only reach a particular page (at the bottom) via one particular route (e.g. People>Names>Charles Darwin; Topics>Ceramics>Porcelain>Coalport Porcelain; Places>Staffordshire>Stoke-on-Trent>James Dudson Pottery)
Fig 4. Conceptual model of 'knowledge-net' database-driven website showing 'knowledge-net' structure of the Darwin Country database-driven website. Each circle represents a page or link and navigating the website from a number of links on the Home Page (around the edge), the visitor can follow a wide range of different routes, exploring the site but 'theoretically' still arrive at the same particular page (in the centre). In practice, the website provides thousands of pages and possible routes which increase as the website grows (e.g. People>Names>Charles Darwin; People>Occupations>Naturalists>Charles Darwin; Topics>Evolution>Fossils>Charles Darwin; Places>Shrewsbury>The Mount>Charles Darwin).
Fig 5. Screen-shot of part of database-driven page where thumbnails of images of the objects related to the subject-matter of the page provide access to larger images provide access to larger images and information about the objects.
Fig 6. Screen-shot of part of a database-driven page of Darwin Country showing text with images about Charles Darwin reached via an initial trail of 'People', the trail of pages followed, links to 'top' categories and 'find', links to categories related to the page displayed and a timeline. The same page could have been reached from 'Places' or via numerous other starting points or trails.
Fig 7. Screen-shot of part of database-driven page of Darwin Country showing example of an interactive map that provides another means of accessing pages about particular places.
Fig 8. Screen-shot of Home Page of administrative part of Darwin Country, providing access to the Content Management System.
Fig 9. Screen-shot of part of input form for 'categories' in Content Management System of Darwin Country, showing how each category is related to up to seven other categories.
Fig 10. Screen-shot of part of input form for 'objects' in Content Management System of Darwin Country, showing how each 'object' is related to up to seven other categories. This also shows how html may be added to allow scientific names to display on the website in italics.