Generally, the masses could not move around the county at will. Nor did they have the desire so to do. Even, much later, this could still be experienced. In my twenties, I used to work in London but spent many weekends in North Norfolk. Sat in a pub one weekend, talking to an old timer, he responded to the question, “… have you ever been to London?” “No, never. Why would I want to? I have been to Norwich three times in my 80 years. That was enough, I don’t need to go any further.” To put that in perspective Norwich was just 15 miles away with good transport links.
Turn the clock back 150years or so, and transport links where not good. The world's first recognizably modern inter-city railway, the Liverpool and Manchester Railway (the L&M), opened its railway in 1830 and proved to be successful for transporting both passengers and freight. In the late 1830s and early 1840s, Railway Mania waned for a short period, but the transformation had begun and was to become an unimaginable vehicle for change. Part of the drive was for the faster movement of bulk freight to help fuel the ongoing industrial revolution. Replacing the canals, the main transport arteries of the time. The Turnpikes had played a part in improving road transport and reducing journey times, but they cost money to use them, and declined with the growth of the railway. Horse drawn carriages were not a normal mode of transport for the poor. Carts, perhaps.
So, what has this got to do with family history and records. When you find a record of a marriage between two people of the same name as those being researched, are they the correct people? How far have they travelled from previously known locations? I plot the potential route in google maps using directions, with the mode of transport being walking. This gives some indication of the possible journey time, without access to advanced transport. How could they have met? Proximity is therefore important to me. However, that is complicated by the enclosure acts and the agricultural revolution. The populous still did not travel at will, but migration had started. Rural became urban, as people flocked to the towns.
Some understanding of this constraint on movement and drive for movement helps with the recognition that two families with the same parents names and dates, and very similar collection of children names and dates, but with locations in Devon and Dorset are in fact different, despite the similarities.
This is why places are so important in my research and recording. Why I spend so much time finding the places on both current online maps and historic maps.
Time and space are interwoven, and are significant and important.
Places and how to address them are a particular interest to me. Structure, data integrity, efficiency of input, information retention and retrieval, and ease of assimilation of information.
Mapping and Geo Positioning is also a keen interest, and manifests itself in travel and the promotion of BIM in the Construction Industry. Another long standing interest is organisational classification, which started whilst I was at school with Carl Linnaeus et al Animal Kingdom classifications. Structured Data.
Places nomenclature within a database environment is a complex issue. It has to cover changes of names, spellings, and boundaries over time, together with significant changes in size. Difficult in a single flat file or list.
I have been working on TNG for a while, and decided it was time establish my Site Conventions for Places and to seek help in the formulation of them. Below is how that started
You find a record and often it is the registration district or parish that is the place.
Then you get more information about the person, and you start to get information about the actual place, which can often appear to be different, and therefore possibly not a match.
Then you look up the Reg district and parish coverage, and find that the place is within the bounds, either current or historical.
One way is to record the most detailed and change the granularity settings, 'Address, Location, Town/City ...', but that then loses the important Reg district and Parish details. The address string and parish string are different so don't lend to being combined. The Lat Lon deals with the actual Geo Positioning but not with the different forms of data.
I tend to use the name of the place as it was at the time the event occurred. Ecclesiastical parishes and county boundaries have changed over the years. Modern English civil parishes often don’t cover the same area as an ecclesiastical parish of the same name. I think it would be a logistical nightmare to track those changes.
My first thought was to have a dataset of the English Registration Districts, with a subset of Ecclesiastical parishes, However, as they appear to change quite frequently that would indeed be a challenge. It is not beyond the capability of modern computers, but unless there is open source data to link into it would, as you say be a logistical nightmare, to create and maintain such a dataset. An alternative would be to have two or three place fields in the database, one for physical location, and another for the church organisational location, and perhaps the third being for civil organisation. All potentially have different answers for the same place. Sorts and finds could also be done on all of the different separate forms of place. Dependant on the date of the event, the town of Christchurch is recorded as being in the county of Hampshire or Dorset, to cover the historic element of the challenge.
My first decision was to ignore the 1974 UK Boundary changes (and later modifications) which eliminates 90% of the problem. My primary resource is the draft document published in 2006 by Brett Langstone which identifies for each civil Parish and Township in England and Wales the Registration districts used between 1837 and 1974. This was intended to be used in the re-organisation of the GenUKI website but an alternate method was used. Where ancient or modern ecclesiastical Parish names were different I use the post 1837 name as the address but identify any alternate names used in the Citation and Source details. I always use the true Country name (England, Wales etc) rather than the Sovereign Name United Kingdom.
A basic tenet of genealogy, including in the advice of early experts such as David Gardner & Frank Smith, on whom I cut my teeth, is to enter the name of the place as it was called at the time of the event.
Thanks for all the comments. I also use the place name at time of event and always include the county and country at England, Wales level. The Lat Lon and how that displays on a map can be the visual assimilation of the current location names. Historic and current considerations are not the main point thought. Of greater interest is the difference between Geo positioned and organizational addresses and retaining as much found data as possible. Geo positioned examples include Lat Lon, post and zip codes and what3words among others. These are relatively simple and constant single point or generally small area locations. Most organisational addresses are hierarchical and are often less definitive. The string of a postal address is different from a Parish / Reg Dist or political / administrative boundary. The records we rely on are often presented by parish and church. Often a village or town has the same name as a parish, so a simple address does not define which it is. This can cause a degree of confusion. If a user wants to pull all the people born in a specific parish together, or all marriages registered in a specific Reg Dist, is there a way to do that?
You have to decide which is more important to you, the reporting as described above, or the ability to automate Geocoding, in the case of the UK they are not compatible. TNG does allow the Place and the Geocode Location to be different formats, but it is a manual process. You should also consider that UK postal addresses have not included County names in the last 20 years, and when they did the Postal County boundaries did not match the County Council and Borough boundaries none of which have a direct relationship to Registration Districts.
That is exactly my concern. I think I have a work around. Use place as intended with BMD events and add custom events for registration of BMD, and store the information on the record in that 'place'. Similar for census records, but possibly two as there are three address types in the Census data. Omit the Sanitary District. Gets more difficult with the Tithe Apportionment as that is often Tithings and Hundreds. This does create a lot more work, but retains all the data from the Source Record.
Well, before adopting a specific schema and convention there are some other considerations to take into account.
In general data management I have for a long time been a proponent of type it in once, use it many times. This was not just for the efficient use of data, but also the integrity of the data. Whilst when we can easily read St Mary’s, St. Marys, Saint Mary, and many other variations, as one place, to a simple computer, they are all different. The worst culprit is the trailing space, as it is almost invisible. A search for St Mary’s, or a list of same does not find all the relevant places. There are ways around this by using multiple variations of the string or using the computer to apply some fuzzy logic. Either of these could either leave out relevant or include irrelevant results, and thereby is inherently prone to errors. By far the most effective and efficient way is to eradicate the problem by using consistent data, input once and used in both the add new data and to filter datasets.
The default position within TNG is to have places created in each tree. That can result in both duplication, and divergence of data. However, I recently found that at a flick of a proverbial switch I could change to a global Places List. Excellent, I though. That will improve efficiency of data management.
However, I did not research my action properly.
I had assumed that the structure would be a central list with flags to respective trees, in a one to many relationships. That assumption was incorrect it seems. The process of changing to global did warn me that all the trees would be merged, which it did in seconds. Success I thought. However, on looking at some of the operations, such as the place map, a significant functionality had been lost by combining the various tree’s place lists. The opportunity to select separate trees to map places had disappeared. This is neither a complaint nor a suggestion, just an observation, and sharing of my error.
I had not done enough research and had made a mistake. I started to reverse it by flicking the switch back to individual place lists, each associated with a tree. I created a report to list all the places in the database, just short of 500. I used that report to reattach the places within the trees to the appropriate events. It took a while. The silver lining was that I had a greater proportion of by addresses with Lat Lon at the end of the process.
The next consideration was the Place Level. A dropdown list offering the default 6 options of ‘Address, Location, City/Town, County/Shire, State/Province, Country.’ I had been working with this level of definition of awhile and found that I did wish for more options to create greater granulation. TNG has the ability for users to create and share Mods. This time before launching into anything I searched the Mods. I found one that increased it from six to ten options. Then a second called Google Maps 15. mod replaces the 6 default Google maps place levels with a range of 15 levels. The default list is ‘Address, Street, Location, Hamlet, Parish, Town, City, Township, Area, County/Shire, Province, Region, State, Country, Continent.’ I found a way to customise this list in the config file and settled on Address, Street, Location, Hamlet, Village, Parish, Suburb, Town, City, Conurbation, Area, County/Shire, Province/Region, State, Country. Where each option is generally larger that the one before. Hence region is a part of a country not the region used by UN classification as a collection of countries. ‘At Sea’ is not catered for, and nor is ‘Somewhere on the continent of (insert name here)’. The use of parish in this context, to maintain the rationality of the list, only refers to urban parishes, which are a division of a town, and not to the larger rural parishes which contain many settlements.
The first pass of Place Level is based on the map that can be displayed as the Place Level is set.
The objective of recording places for me, is not just about an event with a place, but how populations developed and moved.
Trying to do this has led to perhaps an over complex convention. Time will tell if it is achievable, and maintainable, together with how well it delivers the objective.
An option is selected from;
Address, Street, Location, Hamlet, Village, Parish, Suburb, Town, City, Conurbation, Area, County/Shire, Province/Region, State, Country.
Parish is only used for part of an urban environment where it is smaller than a following option. Options will, over time, be indicative of the place at the time of the event.
The list of places will be linked to each individual tree and consistency maintained in the subsequent list in ‘All Trees’.
Places are ultimately defined by the Latitude Longitude single point data. Areas are not specifically catered for. The point representing an area, is either the defined point to represent that area or the name of the area. For instance, Sopley representing the parish of Sopley as well as the village of Sopley.
I have decided not to try to comply with either a 4 or 5 position naming standard, or Geo Coding conventions to aid automatic completion. I consider that such compliance would limit the ability to represent the historic data with the same level of granularity as the records provide.
I use the name of the place as it was at the time the event occurred, and as recorded on the record. Where possible, I review the source information, not just relying on the transcript. If further information becomes available, I add that granularity to the address. For instance, if the first information for place of birth is the town, but subsequently I find out the house of birth, the address will be amended appropriately. With the source of the two pieces of information recorded as associated with the event.
All addresses include county and country where known.
For an actual place, I have used an ordinary address format, excluding Post/Zip Code, which is not strictly a UK postal address format, but something of a descriptive hybrid;
House Name/ Street number, Street, Tithing/Hamlet/Village, (near) Post Town/Town, County, Country
with as much information as is actually stated in the respective event record, with some exceptions.
Implied county and country are added if absent, without added brackets (), based on the date of the event.
If the data is from a Tithe Map the Plot number is recorded in the notes, but transformed into an address for the text.
To aid assimilation of information additional place location may be added, with the prefix near. E.g. Millbrook, Hampshire, England could become Millbrook, (near Southampton), Hampshire, England. That will differentiate it from the later address when Millbrook has been absorbed into Southampton as a suburb, and the address would be Millbrook, Southampton, Hampshire, England.
Different addresses are acceptable for the same place. E.g. Christchurch, Hampshire, England and Christchurch, Dorset, England, are both the same place, but at different times. Southamptonshire is the same place as Hampshire.
Abbreviations are not used as they may not be universally known. E.g. Hants is not used for Hampshire.
Where the only place information is the civil parish or registration district, that or a representation of it is used in the normal address, until better information becomes available.
For the ‘Civil’ Address, this may include Ward, Parish C, Registration District, Parliamentary Ward, followed by County and Country.
Each element will be prefixed by the element designation, followed by the name. Excluding County and Country. This would differentiate it as a civil area place in the place list as opposed to a point.
These are areas, not point locations, but can have Lat Long assigned as a representative point.
Events such as the registration of a Birth is recorded as a separate event to that of birth such that the data regarding the registration district is retained, irrespective of additional information being obtained for the place of birth.
Sanitary and Poor law areas are not currently recorded.
Tithings and Hundreds are not currently recorded, but could be incorporated into the Civil Schema if necessary.
For the Ecclesiastical address, this may include the Parish E, County, Diocese, Province E, Country.
Each element will be prefixed by the element designation, followed by the name. Again, excluding County and Country. This would differentiate it as a Ecclesiastical area place in the place list as opposed to a point.
These are areas, not point locations, but can have Lat Long assigned as a representative point.
