The Digital Domesday
Battle Conference, Glasgow, 29 July 2002
The Digital Domesday, published by Alecto Historical Editions, now joins Katharine Keats-Rohan’s COEL database and John Palmer’s Domesday Explorer to facilitate the recovery and manipulation of Domesday data. Superficially, all three programs are similar. In use, however, they are very different historical tools. COEL is most clearly set apart. It is at once less than a Domesday database and considerably more. Its concerns are exclusively the people of Domesday Book, and it provides documentation of them and their descendents down to 1166 and beyond. It is primarily a database and prosopographical research tool with only a limited number of datasets from Domesday Book. The Digital Domesday is more directly comparable to Domesday Explorer. Both exclusively concern themselves with Domesday data. Nevertheless, the Digital Domesday differs considerably in content, design, and use.
The Digital Domesday is the first database of the whole of the Domesday text, embracing both Great and Little Domesday. It is based upon the Alecto edition, and the critical apparatus that supports it, published between 1985 and 2000. The text itself is presented in three interrelated and linked forms. First, there is a facsimile. In 1984 the then keeper of the Public Records invited Alecto to photograph the manuscripts of Great and Little Domesday which had been taken apart for rebinding. Plates were produced at full size of the Great Domesday manuscript and printed with a continuous tone process to create a state-of-the-art facsimile. At the same time ten by eight inch transparencies were produced of both volumes and Little Domesday was subsequently printed from them using stochastic lithography. It is these transparencies that have been digitized for the database. All of the images – over 1600 in all – are available on the one CDROM in a low resolution format. These not only allow easy reading of the text, but when magnified also afford much detail of the processes of writing such as changes of pen, prickings, and rulings. A full set of high resolution images are optionally available on a further four CDROMs for a more minute examination of pen strokes and the like.
Secondly, Abraham Farley’s transcription is provided. This is the first edition of the text published by order of the King George III in 1783. A fount called Record Type was specially designed and cast for the project in order to capture as closely as is possible in print the variations of the manuscripts. Like them, it therefore requires an understanding of the system of Latin abbreviation used in Great and Little Domesday to read it. Nevertheless, it is a remarkably accurate, if not completely error free, transcription and is useful for confirming readings where the user is unsure. A button brings up the relevant page when needed.
Thirdly, there is a translation which mirrors the form of the facsimile and is set side by side with it. For Great Domesday it is based on the translations published county by county in the Victoria History of the Counties of England produced from 1904 onwards. Between 1985 and 1992 all were revised in the light of modern scholarship and standardized. A new translation of Little Domesday was commissioned in 1999 to match. Throughout the evidence for the identification of place-names was re-examined and for the first time personal names were rendered in a consistent form. The conventions used are set out in the Information section accessible from a button on the opening screen.
Technical terms unfamiliar to the general reader are glossed throughout the translation in pop-up boxes which are accessed by clicking on the highlighted word. The rest of the critical apparatus is accessed from the opening screen. The scholarly introductions to all thirty-five of the volumes of the County Edition of the text are found in the County Essays. They provide not only a commentary on the various sections of the text but together also a comprehensive account of Domesday England. The series of Domesday Studies provides more technical analyses of the text and the various subjects that it informs along with a general introduction. Finally, there is an extensive Bibliography incorporating David Bates’ Domesday Bibliography of 1985 in a list of several thousands of the most significant general and local Domesday studies up to early 2002.
At the most basic level text and articles can simply be read; each section is browsed from the opening screen. However, the Digital Domesday is no mere collection of disparate elements. The mass of data is brought together in a search engine that has been designed to maintain the integrity of the text. Despite the business-like tone of much of the Domesday account, the data are to a greater or lesser degree indeterminate. Context is everything. Broadly speaking, there are two possible strategies for dealing with the problem. The first can be characterized as ‘exhaustive tagging’. For each item of Domesday information codes (invisible to the user) are inserted into the text to convey its every nuance. The record of a single villan, for example, must clearly be identified as a record of a man. But equally it could represent a type of tenure, an assessment, a status, and several other relationships. Different codes can be used to indicate all of these meanings.
Once the whole of the text is coded in this way, closely focussed interrogations of the database can be made. For this reason Domesday Explorer is a very powerful analytical tool. However, the strategy does have its drawbacks. Tagged databases can be difficult to use. To exploit them properly and to the full all of the codes must be thoroughly mastered. For anything more complex than a simple search such databases are essentially command-line driven specialist systems: the user has to know the data and the program to work them. More fundamentally, tagged data are interpreted data; what is coded and how it is coded is ultimately dependent on the interests, understanding, and thoroughness of the designer.
For transparency and ease of use the Digital Domesday has adopted the second strategy. It is based firmly on the text and has been designed to allow it to speak for itself. Ideally, for complete integrity the Latin of the manuscripts would be used. There is, though, no machine readable text and so in the Digital Domesday primary searches are made of the standardized translation. A glossary of terms indicates how various words have been rendered in English. Tagging has not been entirely eschewed. All place-names have been coded as such and so every reference to the same place can be rapidly pulled out of the text. Likewise, personal names have been similarly treated to aid the identification of the lands of individuals. Both of these functions can be accessed from the Indexed Search button on the opening page which opens a list of all the names in Domesday Book.
The bibliography has also been extensively tagged. The Bibliography Search button opens up a list of subject headings. A county bibliography for each Domesday shire can be generated in this way. Publications that relate to more than one shire are cross-referenced and so will automatically appear in all the relevant lists. Other entries will be found for topics such as agriculture, archaeology, castles, charters, the church, demography, economics, feudalism, knights, land tenure, manuscripts, settlement, slavery, and women. In total there are eighty-five different headings.
Tagging, however, is only a limited aid to database navigation. The main way in which the mass of information is accessed in the Digital Domesday is by free text searching. Every occurrence of a word or phrase (including all the tagged information), or combination of the two, can be identified in the text and critical apparatus from the Free Text Search screen by simply typing it in. Technicalities are kept to a minimum. A simple series of words will return every page on which they appear regardless of position. But if double inverted commas are put around them the relevant phrases only will be found. For greater precision there is a ‘proximity’ function. So, ‘salt # #20 eels’ will find only instances in which the word ‘salt’ occurs within 20 words of ‘eels’. Wild cards can be employed if need be. Thus, ‘wom?n’ will find references to ‘woman’ and ‘women’, ‘1 # villans’ all references to 10-15 villans. For the more anorak-minded, the Boolean logical operators of AND, OR, and NOT are supported. Otherwise, there is nothing more to master. It is a simple matter to formulate very precise searches and be confident that you are getting exactly what you are looking for.
By selecting the relevant options a search can be confined to a specific county, or Great Domesday, Little Domesday, the various essays and bibliography, or any combination. On pressing the Search button a list of pages is generated on which the search term is found, and opening a page will highlight the information sought but not isolate it. In searches of the Domesday text in the Digital Domesday every item of information retrieved is displayed in the context of the recto or verso of the relevant folio. The context of the data is preserved at all times.
This design is easy to use and is interpretation free. It is the user who decides what can be searched for and interprets the data when it is delivered. Unlike a tagged system, it cannot instantly produce tight arrays of statistics (although it can, of course, retrieve figures). It is, however, inherently flexible. Until comparatively recently little interest was taken in the formulae of the Domesday text, that is its diplomatic. Tagged systems designed without diplomatic in mind, might require reprogramming to accommodate such an interest. There is no problem with a text-based system. It is as easy to compare the occurrence of, say, the formula ‘he held’ with ‘he held freely’ as it is to find any other item of information. The limits of the Digital Domesday are those of the text and of the user.
©David Roffe, July 2002.