Computers in the History ClassroonAllan Martin & Frances Blow, 1988
A Flexible Approach to Local History Data Bases in the Classroom
Chris F Reynolds
CODIL Language Systems
ABSTRACT: Database packages are used in schools to handle information from well structured historical records, such as census returns or church registers. While such databases can be used for some interesting projects, they present a very narrow view of historical sources. This paper describes the more flexible approach possible using MicroCODIL. The main historical example involves a study of the farms and farmers of the parish of Sandridge, Herts, during the 19th century. Computer files have been prepared from a wide variety of sources, including land tax, poll books, maps, school records, trade directories, census returns, parish registers, wills, a newspaper article, etc. These have been merged into a single data base which allows rapid access to biographies of the farmers, and histories of the farms.
Local history can play an important role in the classroom in allowing pupils to relate what happened to the area in which they live. Setting up a database can allow children to explore the information themselves - and sometimes they can help "collect" the data if the source documents (or photocopies) are suitable. Most existing teaching databases use census returns or parish registers, as their contents were determined by law, have a regular structure, and are reasonably comprehensive.
On the other hand, to limit "local history data" to that which can be organised into neat and tidy tables is rather like teaching art by confining the pupils to straight lines. This is not to decry their value as cross-curricular studies are to be encouraged; some work could easily be seen as us ing local history data to teach simple statistical analysis.
This paper describes how MicroCODIL can be used to set up historical databases which can contain information from a very wide range of documents, each of which may have its ou~ individual structure. The main historical example involves a study of the farms and farmers of the parish of Sandridge, Herts, during the 19th century. Computer files have been prepared from a wide variety of sources, including land tax, poll books, maps, school records, trade directories, census returns, parish registers, wills, and even a newspaper article. These have been merged into a single database which allows rapid access to either biographies of the farmers, or the histories of their farms.
NOTE: MicroCODIL uses the term "knowledge base" to include the sum total of all information about an application. The term "database" is used here when the majority of that information is held in a single large file within the knowledge base ..
MicroCODIL is a byproduct of university based research into human-computer interaction, ulith specific relevance to poorly structured information and open-ended problems. It is designed to allow pupils access to a wide range of advanced information processing and expert system ideas in a classroom context, and can be used as both an aid for teaching information technology and for cross-curricular studies. MicroCODIL is currently being marketed for BBC computers and a special History Pack has been produced. Due to lack of funding it has not yet been possible to carry out formal classroom trials, but initial reactions by the educational press have been very favourable.
3. THE HISTORICAL KNOWLEDGE BASES
Because the MicroCODIL approach is novel the software is provided with a series of example knowledge bases, covering a u1ide variety of subjects and information processing activities. Five of these are of interest to the History teacher, and are briefly summarized below:
KINGS: This offers what seems a simple information retrieval facility on English Monarchs from 955 and history teachers might well use it for that purpose. It can also be used as a historical example for the IT teacher, demonstrating several different techniques for finding the answers.
PRISON: This deals with prisoners held in Aylesbury prison at the time of the 1851 census, and was selected to show how 'pro forma' type data can be handled by MicroCODIL including the use of an interface to conventional BASIC files.
DEVON: At first sight 'lists' of names would seem easy to fit into a conventional database. This example shows that this is not the case, and that the information can be very complex indeed. It is based on a list of Devon MP's from an 18th century register.
INVENT: This is based on 5 inventories from 1669-1709 and shows a hierarchical classification of items (HOUSEHOLD goods includes BEDDING, FURNITURE, FIREIRONS, etc) combined with original spellings. This will help pupils relate the original text to the uses of often unfamiliar items. Thus "FURNITURE (MATCH) = stools" finds "6 stoals", "1 stool", "stools" and "1 closet stoole". A search for "GOODS (STRING) = IRON" finds "FIREIRONS = 1 pair of handirons" in the house, "IMPLEMENT = 1 ironrake" in the ma1tloft, and "STORES = the iron old and new" in the blacksmith's shop.
BOGBOD: This relates to the discovery of Grauballe Man and is based on a text searching example written in Prolog. However it takes the example much further by including "expert system" style facilities to automatically solve the problem, with help texts providing a commentary.
4. A SIMPLE EXAMPLE
Before the Sandridge database is described in detail it is worth looking at a simple example involving a single comparatively straightforward source document. The document is an old leather-bound book (now in the Herts Record Office) listing the Sandridge Parish Officers. Each page is headed with the name of a parish post, and each year the names of current parish officers were added. Figure 19.1 shows the page headed "Stonewardens" which runs from 1859 to 1874.
At first glance the document seems an easy one to handle using a conventional database package, but close examination reveals a number of awkward features. Here is part of this information rewritten in MicroCODIL; underlined items are those which would not easily fit into a simple "one year, two name" pattern:
Basically there are three breaks in the pattern. The first is that in 1865 and 1867 the name of the farm is given in brackets, although everyone named was a farmer. If a conventional fixed format package were used, there wou1d be a tendency to leave off this extra information because it would be inconvenient. This would be a pity as, in any historical documents, a non-standard entry usually means something interesting has happened, and such "odd" information often provides a memorable teaching point. In this case the farmers didn't want to do the job of maintaining the village roads, and so the job rotated round the farms. In 1865 it was the turn of Samuells Farm to provide the stonewarden; Edward Woollatt was the new tenant.
The unusual entry in 1868 shows that the normal allocation of duties u/as carried out, but as a result of a change in the law, a new post of waywarden was created. While entries continued to be made year by year the waywarden post no longer rotated, and it is sensible to represent this by using a range, rather than typing each year in turn. (MicroCODIL. can automatically handle such ranges so no technical problems are created in handling the data in this way, although in more conventional database packages special coding and programming would be needed.)
5. THE SANDRIDGE DATABASE
Sandridge is a rural parish just north of st. Albans, and was chosen because it was already part of a larger study into farming families using a mainframe computer. This meant that a very significant quantity of information was already available. In fact the total volume of information available would have swamped a normal school BBC computer and it was decided therefore to select a subset of information which would just fit on a single-sided 40-track drive for classroom use. The chosen subset was restricted to the farms and farmers of Sandridge in the 19th century and, with a few deliberate exceptions, their families were excluded. A series of extra source documents were specifi~ally included to test the software and the suitability of the approach. The idea is that the database can be used directly, but that it can be used by schools also as a model to set up their own local history database.
The data base is constructed from the following independent source files:
The idea was that each file relates to one document or class of documents, although some files, such as bapt+, contain similar information from other sources. Each file consists of lists of items, as in the example above, using item names appropriate to the source document. Thus tax20 contains items such as LANDLORD and TENANT, office+ identifies people as OFFICER, while bapt+ refers to FATHER, MOTHER, SON, etc. Thus the information is held in a form which can readily be related to the source document, and if photocopies are available, and pupils are of a suitable age, they could be allocated documents to code up. In this case some information came from the mainframe database rather than directly from the original historical documents.
The file other+ contains information from a variety of sources, such as family histories, confirmation records, court records, wills, and newspaper articles. It allows unusual and interesting items to be included which will fire pupils' imaginations. Here is a subset of the file: it includes the shooting of a panther, the conviction of a leading farmer, and a murder, unusual events which will liven up a class using the database, while the more observant might realise that Earl Spencer was Princess Diana' s ancestor.
Constructing the database involves copying individual files to a large single 'database' file, labelled with the year and source file name as appropriate. MicroCODIL is then used to produce a series of indexes, in this case by name and place. Since people can be referenced in different ways a 'PERSON" is defined as a hierarchy of terms. Here is the PERSON hierarchy:
If you are using new data there are bound to be problems that come to light. In some cases it is appropriate to correct the original file, but in others it is necessary to explicitly identify the name or place. In the case of the Sandridge database this is well illustrated by the 1869 trade directory. Pupils are invited to answer the question "What makes you think that the 1869 trade directory was not very accurate?". Here is part of the trade+ file for that year:
The INDEXNAME items have been added because the original document was full of errors, with "George" instead of "Jacob", "Edward" instead of "Edmund", etc. When setting up a new data base such real world problems will need to be attended to.
The complete database is menu driven. Options 1 and 2 list the people and places mentioned on the database, option 3 generates a biography for any named person, while option 4 gives the history of the named farm. Option 5 a11ou's the pupil to define a search of the complete data base or (much faster) on of the source files. Option 6 a11ou~ the pupil to write his own MicroCODIL statements to process the files. Here is a typical biography:
The documentation includes a range of questions which really involve minor historical investigations. For example one question is 'Who died in the Union Workhouse? What other factor suggests that he was not as well to do as the other farmers?' The clue suggests a search for the union workhouse while a search of the biography of the named individual reveals that he paid much lower rates than the other farmers - suggesting a much smaller farm.
The research described above shows that the technology works and that it is possible to combine historical data from many different sources in a stimulating manner. The next stage is to get the system used in a variety of environments from primary school (with the teacher preparing the data) to a real local history research project (part of WEA classes?).
A variety of topics, such as the history of an actual school building. a mill or some local political movement. would add to the scope for further assessment of the technique.
In addition it should be realised that many of the advanced facilities of MicroCODIL, such as probabilities and fuzzy logic. are not used in the current example, and that there is scope for even more interesting approaches to local history once the software is operational on more powerful classroom computers. such as the Archimedes.
Extensive further information on MicroCODIL is given in the MicroCODIL Manual and MicroCODIL and History, which are supplied with the software by CODIL Language Systems Ltd. The most relevant additional publications are detailed here:
Reynolds, C.F. (1988) "Introducing Expert Systems to Pupils" Journal of Computer Assisted Learning 4: 79-92 Gives exa mples of use of MicroCODIL in a number of teaching areas; includes brief description of BOGBOD knowledge base.
Reynolds. C.F. (1988) "CODIL as a Knowledge Base System for Handling Historical Information" in Ratz, S., Computer and Quantitative Methods in Archaeology BAR International Series 446(ii), 1988. pp 425-433 A more general paper, includes example relating to study of the Phipson family, carried out on a mainframe version of the software.
Reynolds, C. F. (1988) "A Psychological Approach to the Computer Handling of Historical Information" Proceedings of the Cologne Computer Conference Includes detailed description of the DEVON knowledge base.
Four family histories have been published. based on computer listings from MicroCODIL's mainframe precursor. These are The Gibbs Family of Aylesbury, The Rendell Family of South Devon, The Phipson One-name Study and The Ancestors of Lucy Ann Reynolds. These are available from CODIL Language Systems.
ADDRESS FOR CORRESPONDENCE: Chris F Reynolds. CODIL Language Systems Ltd. 33 Buckingham Road, Tring, Herts HP23 4HG. U.K.