Updated 6 Dec 2000
WIRKSWORTH Parish Records 1600-1900
Return to Front Page
This is a draft of an article which appeared in the November 1998 Newsletter of the Society of Indexers Genealogical Group.
Deciphered, transcribed, compiled, indexed, formatted and copyright © 2000, . All Rights Reserved.
This is the story of the copying to computer of the parish registers of Wirksworth, Derbyshire covering the period 1608-1837. 70,000 baptism, marriage and burial records were involved, and the goal was to make these available to amateur researchers world-wide, especially those trying to build their Family Trees, and to store the database in a reliable low-tech format for posterity.
The project was begun in July 1995, and has now been brought to a successful conclusion after 3 years and 2,500 hours work. The vital events are stored in a 30 Mb database on a ZIP computer disc, a set of 14 microfiches is on sale, the database is browsable on the Internet and the author's Free Search Service has replied to over 1,000 enquiries from 6 countries.
It was realised from the beginning that a project of this size and complexity could only be carried out using a computer database, copying directly from microfilm displayed on an adapted microfiche reader. Therefore a 486 computer was purchased with "Masterfile Professional" general purpose database software running on DOS, a ZIP disc system for the essential backing-up, and a simple microfiche reader adapted to hold microfilm.
The Genealogical Society of Utah were most helpful in loaning 6 microfilms for use at home. In return they expected 6-monthly progress reports, and a complimentary set of microfiches at the conclusion of the project. These microfilms formed the basic source of the database, but there were a few small gaps in the register record and a two year gap around 1670. Lichfield Joint Record Office were most helpful in providing photocopies of the original Bishop's Transcripts 1662-1701, which were used as the secondary source of the database.
Permission to copy was obtained from the Derbyshire Record Office, the Genealogical Society of Utah, and the Incumbent of St. Mary's Church, Wirksworth.
The event format to be recorded depended on the decision that baptisms, marriages and burials would be combined in a single database. Also that one event must be printed out on a single line. This decision gave greater ease of use when attempting to build a family tree, the raison d'etre of the project. Therefore, the following standard fields were recorded from each vital event: Surname code (see below), Event type (C, M or B), Date, Surname, Firstname, Relation, Placename, Father's firstname (or spouse's firstname), Mother's firstname (or spouse's surname), and finally Comments. Two records were produced for each marriage, husband leading or wife leading, one produced from the other by macro for speed and accuracy. In this way, each partner in a marriage could be located in the same alphabetical listing. Other fields were recorded, such as date of birth, viewable on computer but not on microfiche or Internet for lack of space. Some information, such as age on burial and occupation, was recorded in the Comment field. All surnames were displayed in upper case, and date as yyyy-mmm-dd, which was considered much the easiest to read.
Surn Code Event Date Surname Forename Abode Relative's Names Comments 1988 C 1732 apr 12 WRAGG James Callow son James Mary Farmer 1943 M 1763 jan 23 WILCOCK George Cromford husb Ann FEARN Lead miner 601 M 1763 jan 23 FEARN Ann Biggin wife George WILCOCK Spinster 1187 B 1799 dec 31 MACHIN Eliza Ible wife Peter Joiner
The chief users of the database was foreseen to be amateur family tree builders, therefore it was decided to translate all Latin to English, expand all forename contractions where possible, and use the modern Gregorian calendar throughout. Three main indexes were to be compiled: surname-firstname-date; surname-parent's firstnames-date; and date. Used together on a computer, these three indexes allow powerful ordering and searching and are the most useful for building family trees. To date some 500 family trees have been constructed using this system. Another six indexes are compiled and available, involving placenames, occupations and comments but these are rarely used.
One of the main problems encountered was how to handle variants on surnames, firstnames and placenames. One Derbyshire surname (GREATOREX) was spelled in 43 different ways, 70 more in over 10 ways. Often users have not appreciated the size of the problem before 1800. After several false starts, a system was developed that has proved of the greatest convenience. First, the computer was used to list, in alphabetical order, every different spelling of all surnames (there were 5,600). Then these surnames were put into groups, forming different variants of one basic surname (there were 2,000 groups). Each group was then given a simple code number (e.g. 1 to 2,000). The index was then compiled based on the code numbers. This had the effect of pulling together and listing all different spellings of one basic name, which proved to be very easy and restful to search by eye, on paper or screen. The same idea was used with firstnames (1,500 different with 500 codes) and placenames (760 different with 350 codes).
Originally it was intended to use Soundex or Phonix to group variants together, however it was found they were not successful in dealing reliably and consistently with old Derbyshire names. Instead, grouping was performed "by hand", each name being considered and grouped individually by the author with paper and pencil. This approach was very time-consuming but reliable and consistent. Once codes had been established for all spellings, the computer "bulk edit" facility was used to assign every record its own codes. Some 2,000 codes were distributed among 300,000 names in this manner. On the microfiche, surname code was printed first, enabling the user to go straight to the basic surname after looking up the code in a separate list.
Another problem was deciphering handwriting on microfilm. The main difficulties were style and faintness. The best solution to style was found to be experience, and comparing with other records by the same scribe. Faintness was largely solved by examining the original Parish Registers under ultra-violet light, which had the effect of making the parchment glow and throwing the script into better contrast. A great help was found to be computer cross-matching. As the database grew larger, it became profitable to search for a related but readable entry, which often provided the solution. Some 500 records were solved in this way. The Bishop's Transcripts were consulted with caution, because of their inaccuracies. Another approach was to hire a palaeographer for clear but indecipherable cases, but this proved expensive and still arguable. Finally, a residue of 370 records remained unreadable from 70,000, about 1 record in 189. Surprisingly, several of this residue have been solved by other eager researchers, who have written to the author with their solution.
Of greatest importance is an efficient Error Control System. As the Transcriber proceeds, he must be very honest with himself and insert a "?" character in any deciphered field which leaves a slight doubt in his mind. Without this system it quickly becomes impossible to re-find a doubtful record in the maelstrom of confident records. Honesty is certainly the best policy. With this query tag, it becomes simple to search for, display and review the problem. Particularly stubborn cases can be attacked with increasingly potent methods: BTs, cross matching, UV examination, peer revue, and palaeographer's opinion until the problem is solved or the Transcriber admits defeat. The unsolved field continues to keep its query until an inspired user finally appears with a Rosetta stone.
As the database grew larger, discipline, routine and hard work became important. Enthusiasm and a tolerant wife proved invaluable. "Little and often" became the watchword for typing, ideal for a retired enthusiast working at home during winter. It was found best to type directly into a pre-formatted word processor file, matched to baptisms, marriages or burials, and then import into a temporary database. Inevitably "typos" and misspellings arose, but by indexing on each field in turn and scrolling, these were caused to stand out and could be corrected easily before exporting to the main database. Much labour was saved by typing numbers to represent the commonest placenames. These were replaced with text by using the computer's "bulk edit" facility.
Backing-up became of supreme importance, the possibility of losing the fruits of 3 year's labours could not be risked. ZIP discs were chosen as the ideal way to back-up, since the database with pre-compiled indexes grew to become 30 Mb. Inevitably, the author's computer did crash one dark, wet February afternoon, but only the morning's work was lost because of a rigorous daily back-up ritual, religiously adhered to. Two years' hard work was saved, and suicide unnecessary. Let copiers beware!
After completion last Christmas, and subsequent celebrations, the database was run from ZIP disc on computer making use of 9 available pre-compiled indexes. This enabled locating and displaying any record in about 3 seconds, using any index. By entering the first few characters of a surname, the surname code could be displayed in 3 seconds. By re-entering this code number, all the variants of this surname could also be displayed in 3 seconds, in firstname order. A different index (e.g. parent-ordered) could be displayed in the same time. With this system, it is remarkably easy to work backwards in time from a given individual, flipping between firstname and parent ordered display, tagging relatives, establishing parents, siblings, maiden names, second marriages, burials, wife's family, (assuming the data is there in the first place). Selecting tagged entries and displaying in date order produces a list from which a family tree can easily be drawn. Practice ensures more thought is given to the family tree than driving the computer.
Other databases were compiled, in particular MIs and Directory Tradespeople, and links established between those and the Parish registers. For the first time, a large scale comparison was conducted between Parish Registers and Bishop's Transcripts. Ten thousand events were directly compared, and the error rate of the BTs shown to be surprisingly high. The attitude of the scribe was clearly "who's going to check it anyway?", 300 years before it was.
The PR database on computer was ideal for compiling demographic statistics. Day of the week was calculated from the date, and a sudden and popular switch in baptisms from Sundays to Wednesdays uncovered. Local Mortality Crises were located, when the burial rate over a couple of months increased fivefold or more, and attempts to identify the causative epidemic made from ratios between adult/child and male/female burials. Surprising changes in population, seasonalities in marriages, baptismal delays and a forgotten "baptism festival" were all quantified. Life expectancy at birth and population age structure studies await the time necessary to link a significant number of births and burials. Gaps in days between records were easily calculated, and all gaps longer than a specified limit were printed in date order.
A Free Search Service was offered via e-mail and "snail-mail", advertised on genealogy bulletin boards and in magazines. Before establishment of the Web-site, a search was done on the surname requested, ancestors of the named individual being identified and tagged, attached to e-mail or printed out with accompanying comments, and sent or posted. This proved popular and a fertile way to meet interesting people in many countries, and a superb way to get a feeling for local history.
Because of the care and effort involved in deciphering the Wirksworth Registers, and the fact of its successful conclusion, it would not surprise the author if other transcribers were deterred from repeating the project for this parish. Therefore thought was given to preserving the results for posterity. Magnetic recording, CD-ROMS and computers were all rejected, because of the speed of evolution of hardware and software, and the correspondingly short lifetime of any chosen system. Magnetic domains may survive for 50 years, but the software and hardware to read them will probably be unavailable after 15 years. Printing in book form was also rejected, on grounds of size and cost. The 70,000 lines of print required, with font small enough to fit 10 fields of data into the width of an A4 sheet, would demand 1,000 pages of A4 paper and a spine width requiring a very expensive hard back binding.
The solution lay in the microfiche process. By printing onto American Legal size paper (8.5" x 14"), it was possible to reduce 60 pages onto one microfiche, and shrink 70,000 records of 10 fields each onto 13 microfiches. More important, microfiche are robust, cheap, need minimal storage space, and require a very low-tech data-retrieval system. The microfiche reader consists only of a bulb, a lens, a mirror and a screen, and is likely to be available unchanged in 200 years time. One microfiche costs about 30 pence, enabling the author to sell the complete database, together with 100 pages of related data, as a set of 14 microfiches at a cost price of 6 pounds including airmail postage. Such a low price has resulted in the sale of 1,000 microfiche world-wide in 3 months. One problem has been the collection of such small amounts of money from foreign countries, because the currency conversion fee overwhelms the cost of the product. A sympathetic bank manager has been a great help.
For truly world-wide dissemination of the database, and ease of updating, the Internet was chosen. This required a Modem, a Server, a web-site, understanding of a simple computer language called HTML, and about 4 week's work. Existing database records were first compressed to fit onto a computer screen more easily, which meant abandoning the columnar layout but at the same time making the database much more resistant to large scale importing to database by unknown copyright breakers. A simple macro can convert columnar to compressed format. The occasional line-wrap could be tolerated. Then the database was divided into about 300 sections of more manageable size to speed transfer over the World Wide Web. Each section was filled with hundreds of "tag" characters, using the exchange function in the Word Processor software, enabling the Internet to display the database in the required format on computer screens thousands of miles away. Finally, some software called "FTP" enables errors to be corrected and more data added in a couple of minutes. This allows the web-site to be kept 100% up-to-date with the author's on-going research and with comments from browsers world-wide. The results can be viewed on the Internet at:
The author has found the copying of the Wirksworth Parish Registers to be tremendous fun, involving old handwriting, dog Latin, history, statistics, travel, computers, the Internet, advertising, sales, a network of friends around the planet and a product that should prove useful to many people for generations to come. Gentle reader, consider copying your favourite parish registers!
If the reader has any queries about this article, please contact the author:
Deciphered, transcribed, compiled, indexed, formatted and copyright © 2000, . All Rights Reserved.