Dr Jasmine Allen, with contributions from Dr Joseph Spooner and Dr Martin Crampin
Great Britain is home to many tens of thousands of stained-glass windows. They are found in a variety of architectural contexts – not only in our places of worship, but also in civic, educational, recreational, and domestic buildings. In comparison with other major forms of public art, this huge body of stained glass (and particularly that made after the Reformation) is little known, poorly documented, and under researched. Yet many of these windows have enormous cultural, historical, and artistic value, both for their local communities and for the nation.
Without basic information about our stained-glass heritage – locations, subject-matter, dates, creators, commissioners – we cannot properly understand its context, interpret its meaning, or assess its significance. The vulnerability of stained glass in churches remains as pressing an issue today as it was fifty years ago during the first great wave of Church of England church closures.1 Closures are now a concern across multiple denominations, as was highlighted in the recent event on the future of church buildings organised by the Victoria and Albert Museum in partnership with National Churches Trust, Churches Conservation Trust, and The King’s Foundation.2 This is coupled with the news that the Church of Scotland will close around 30% of its churches by the end of this decade.3
In the face of such major challenges, the recording of stained glass remains as vital as it was when the Corpus Vitrearum was established in the aftermath of the Second World War. Even basic information provides a record that can be kept for posterity or used to help develop a case for the retention (or, conversely, the removal) of a window from a building under threat. Furthermore, by providing enhanced levels of information in publicly accessible databases, we can encourage the wider appreciation of this art form, as well as help to contextualise and interpret Britain’s stained-glass heritage for global audiences.
The Corpus Vitrearum (Great Britain), a British Academy Research Project, has pioneered the cataloguing of medieval stained glass in England, both in print and online. Yet, the post-medieval period, which accounts for the vast majority of extant stained glass in the British Isles, has been largely and impressively recorded by dedicated and knowledgeable individuals, many of whom have set up databases and websites to make this information freely available. Together, these open access resources have enormous value as the first port of call for researchers. Existing data however is currently presented in wildly different formats across different databases, to varying degrees of detail and accuracy, and held on disparate websites managed by individuals. This means that the future of these resources collectively is highly insecure.
An Invaluable Collection of Resources
It was against this backdrop that in May 2024 representatives from a total of thirteen different online resources dedicated to recording information on the stained-glass heritage of Britain and Ireland gathered in Ely, Cambridgeshire, for a workshop on ‘Recording British Stained Glass: The Future of Databases’ with a view to discussing these challenges (Fig. 1).
Organised by The Stained Glass Museum in partnership with the British Corpus Vitrearum, the one-day event was generously funded by the Paul Mellon Centre for British Art and the Worshipful Company of Glaziers and Painters of Glass. The purpose of the workshop was threefold:
- to bring together for the first time disparate groups of people who own, manage, or contribute to unique public resources for researching stained glass in Britain and Ireland;
- to gain an insight into the collective progress that has been made in cataloguing and researching stained glass, and to identify gaps and possibilities for future recording;
- and to discuss together ways in which we can ensure that this data remains available to future generations, identifying opportunities and challenges, and addressing important questions about the future of these databases and resources.
Prior to the event, a questionnaire was circulated to find out more about how the resources run by the participants came about, and to get a sense of their scope and coverage, how they are funded and managed, and whether there were any plans for development, preservation, or archiving. The questionnaire also requested numeric data about the numbers of entries and images, as well as some basic technological information (website platforms, software, taxonomies, metadata standards).
Analysis of the responses provided an overview of the current stained-glass surveys in Britain and Ireland. It became clear that whilst these resources present data in different ways and to varying levels of detail, together, they present an invaluable resource for our stained-glass heritage. Collectively, these resources cover a broad chronological span, from the earliest surviving in situ medieval stained glass (1100) to the present day, and a broad geographical area. Many of the websites are attempting a comprehensive recording of a defined geographical region or county within the British Isles,4 yet certain geographical areas – notably Wales and the south-west of England – are far better represented than others (such as Scotland and the North of England).
Opportunities and Challenges
Perhaps unsurprisingly, the data focused overwhelmingly on churches, predominantly Anglican ones, and the recording of stained glass in secular buildings and places of worship in the care of other Christian denominations is nowhere near as comprehensive.5 Other areas of opportunity for development include being able to make connections between stained-glass windows in different locations (e.g., dispersed glass, or windows made from the same cartoon or design). Existing datasets also rarely include information on preparatory drawings and cartoons, or about the conservation of a stained-glass window, and only a few websites include references to archival and printed sources that would enable the researcher to find out more.
Furthermore, these resources are currently held on disparate websites that are not always easy to find, unless you already know about them, or are redirected via another central website (for example, those of the British Society of Master Glass Painters or The Stained Glass Museum). Yet, these websites already have large numbers of users, ranging from individual researchers and architects to parish churches and communities carrying out research for their own interest or for heritage projects. The use of such sites has grown significantly since the inception of these online databases in the 2000s.
Discussions at the Ely workshop continued in small and large groups throughout the day. It was clear that although these resources may have been set up for different reasons, those who own, contribute to, and manage them had a common interest and appetite for sharing information about our stained-glass heritage. The groups also shared the challenges they faced, including migrating data, implementing structural changes, and making updates to keep sites secure and accessible.
Sadly, it also became clear that the long-term future of these resources is uncertain. Most of these websites are projects run by individuals, and reliance on individuals makes the risk of an uncertain future greater.
Priority Areas for Development
Given this uncertain future, the archiving and preservation of extant data was identified by attendees as an immediate priority. The practicalities of collaboration and potential aggregation of data were discussed, including the need for funding, maintaining, and facilitating volunteers; the advantages of organisations hosting sites; and plans for maintenance and sustainability. Attendees were generally open to the idea of contributing their data to a central resource or repository in the long term but also felt it was necessary to retain the vibrancy of individual, often localised projects, making it essential to consider the relationship between any central resource and its federated sites. Furthermore, there was the question as to what format a central resource would take, and who would have responsibility for it.
The primary object within almost all the databases surveyed was the window itself, rather than an image of a window. This is one of the main points of divergence between the Picture Archive of the Corpus Vitrearum (Great Britain, https://www.cvma.ac.uk) and many other online resources, which have been developed principally to record post-medieval stained glass in situ.
Conversations were had on the needs to develop data standards, which are useful for enabling interoperability, and to agree minimum data recommendations, or ‘essential categories’ for standardising data. Currently, few sites conform to nationally or internationally established data standards. Core data – for example, artists and makers, dates, locations, and (ideally) subject-matter – can be found on most of the websites, but variations in vocabularies mean there is a lack of consistency, resulting in a wide variation in the way that names, dates, and places are presented. Inconsistency in the data within individual websites is a particular impediment for accurate searches – and about half of the websites discussed cannot be effectively searched in their present form.
Not all the databases represented at the workshop include images. The importance of good-quality images accompanying data entries was recognised, although it was also acknowledged that the hosting of large numbers of high-resolution images on websites brought additional costs, as well as questions around copyright and licensing.
Perhaps, most soberingly, it was acknowledged that most of these databases do not have any plans for the long-term archiving of their data. This is one of the main differences between those websites run by organisations, which have benefitted from funding and are more likely to have a strategy for the future, and those that are run by independent individuals using their own funds.
The Digital Unit of the Corpus Vitrearum International
In 2022, the Corpus Vitrearum International established a Digital Unit, open to the fifteen national committees of the Corpus. The Unit is responsible for recommending digital standards in relation to the study of stained glass. It focuses on digital image archives; the creation of standardized and structured metadata; digital publication; and ensuring the accurate digital recording of historic stained glass for future research and publication. Its priority is to ensure that data is findable, accessible, interoperable, and reusable. To this end, the Digital Unit has authored a working paper on ‘Matching Metadata: Recommendations for the Digital Resources of the Corpus Vitrearum International’ (2024),6 which is available open access [https://doi.org/10.5281/zenodo.10829493] and provides useful recommendations on matching metadata for stained glass recording projects. Two representatives of the Digital Unit, Sarah Pittroff and Jonatan Jalle Steller of the Digital Academy at Mainz Academy of Sciences and Literature, outlined to attendees at the Ely workshop some of the projects taking place internationally to connect existing German and Swiss databases of stained glass and to agree international metadata standards.
In order to move towards the interoperability of stained-glass data specifically in Britain and Ireland, a small working group met several times following the Ely workshop to explore how these resources might enable the retrieval of metadata from across websites. The working group has subsequently made a number of suggestions for improving metadata standards and the ability of search engines to find data on the sites (such as embedding of RDF headers in webpages using schema.org and creating more structured data).
A new web-based cross-site search engine proposed by database developer Nigel Callaghan in collaboration with Martin Crampin (Fig. 2) (who created the online Stained Glass in Wales catalogue [https://catalogue.stainedglass.wales/]) would enable users to search across multiple resources and combine results from different databases, using embedded APIs. This would present data on stained glass from different websites in single combined searches, with links to further information and images on their respective sites. It would also provide better searching of sites that are not currently fully searchable, and be able to generate periodic archivable backups to assist with concerns about data preservation.
Metadata and the Possibilities of AI
One of the main obstacles to the provision of data about stained glass is the time needed for fieldwork and data entry, particularly the knowledge required to identify the maker of a window. With the vast amount of stained glass recorded on larger online platforms such as Flickr (a commercial site) and Geograph (a not-for-profit enterprise), and in key listings, such as the Pevsner Buildings of England series and BSMGP lists, some stained glass recording can be carried out at a desk, but verification of sources and photography still needs to be carried out on site.
Another recent event ‘Modern Stained Glass-Metadata-AI’, held in Münster, Germany, in September 2025, highlighted how the creation of metadata could be done with a degree of automation. This workshop brought together people from across Europe to discuss how metadata is used in inventory and recording projects, and how AI might also be harnessed to facilitate the development of such resources. Presentations demonstrated how AI models can make creditable attempts at subject indexing from images, and create descriptions using machine learning, provided that reliable high-quality training data has been used to train the model. Although these models may struggle to distinguish windows by makers with the same name, e.g. Powell Brothers of Leeds from those by Christopher Charles Powell of Highgate.
Larger, funded research projects have tended, from the outset, to implement specific data standards and principles and make plans for archiving of the data. One research-driven international example is Vitrosearch [https://vitrosearch.ch/en], a joint initiative of the Vitrocentre and the Vitromusée Romont, the national glass centre and museum in Switzerland. Notably, these larger projects have received funding from institutions at national and local levels, higher-education institutions, and denominational bodies (as with the work underway at the Faculty of Catholic Theology, University of Münster, funded by the Catholic Diocese of Münster to record all the fittings and furnishings in the churches within the diocese).
Another method for data gathering is to open up resources to crowd-sourcing, where members of the public are able to provide information and images to help improve the quality and quantity of searchable data. Although this does not remove the need for trained cataloguers, the latter’s efforts in editing and verifying data may go much further when making use of such crowd-sourced assistance. One of the delegates at the workshop, Anique de Kruijf, from the Museum Catharijneconvent, Utrecht, Netherlands, spoke about the Kerkcollectie digitaal project [https://kerkcollectie.catharijneconvent.nl/], which is run along these lines to create digital inventories of all church and convent buildings in the Netherlands.
International projects continue to provide innovative examples of maintaining data accuracy and recommended data standards, whilst preserving and making information on stained glass heritage more available to various audiences. The recent archiving of the Picture Archive of the Corpus Vitrearum (Great Britain) with the Archaeology Data Service (ADS) has ensured that this resource is preserved for the future. The British Corpus will have the opportunity to expand the resource, and its relocation is welcome news [https://www.vidimus.org/news/news-from-the-corpus-vitrearum-picture-archive/]. Looking ahead, ensuring the preservation of other independently managed resources is a pressing concern. Developing existing resources so that they are more easily interoperable will require significant funding as well as a desire for collaboration, but it is clear that there are many opportunities for these resources to maximise their use, both now and in the future.
Footnotes
- Marcus Binney and Peter Burman, Change & Decay: The Future of Our Churches (London: Studio Vista, 1977). [↩]
- ‘Great Expectations: Churches: See them. Save them. Support them’, 21 October 2025.[↩]
- https://www.scotlandschurchestrust.org.uk/blog/scotlands-closing-churches/ (accessed 30.10.25).[↩]
- In this sense, they attempt the sort of survey for glass that Pevsner’s Buildings of England series did for architecture.[↩]
- Only 40% of existing online resources recording stained glass in the British Isles featured other Christian denominations and places of worship. Only 50% included secular spaces.[↩]
- Steller, J. J., Pittroff, S., Knorr, M., Lecocq, I., Rodrigues, A., Spooner, J., & von Wartburg, K. (2024). Matching Metadata: Recommendations for the Digital Resources of the Corpus Vitrearum International (1.0.0). Zenodo. https://doi.org/10.5281/zenodo.10829494[↩]