* Electoral Register v Census v 1939 Register

AS allows faster and more convenient creation of source records for Family Historian.
Post Reply
avatar
ColinMc
Superstar
Posts: 458
Joined: 17 Jan 2019 11:35
Family Historian: V7
Location: Edinburgh

Electoral Register v Census v 1939 Register

Post by ColinMc »

If I try to enter an Electoral Register entry by selecting an individual, for 1939 when I already have a 1939 Register entry for that person, AS tells me that I already have an entry for that person.

I checked that a similar message is given if the 1911 census is in existence and a 1911 Electoral Register is selected. I have not tested if the other census dates act the same.

I also note that in the Facts Tab of the Property Box, my Electoral Register entries are referred to as Censuses. I can't see anywhere in Options to change this, but it may be something I have set. If so, how do I change it back.

Thanks
Colin McDonald - Researching McDonald, McGillivray, Tait, Rountree families
User avatar
tatewise
Megastar
Posts: 28341
Joined: 25 May 2010 11:00
Family Historian: V7
Location: Torbay, Devon, UK
Contact:

Re: Electoral Register v Census v 1939 Register

Post by tatewise »

In the Tools > Census Templates... you can choose Census or Electoral Register and the default option for the Fact: setting is to Use the default fact which is chosen in Tools > Options > Census settings and thus cause the 'conflict' you are seeing. However, the other options allow a specific fact to be used. Therefore, for Census entries it could be set to Use census fact and for Electoral Register entries it could be set to Use residence fact and avoid the 'conflict.

That is covered in the Help for Census Templates, but could perhaps be highlighted somewhere else because that 'conflict' is going to arise sooner or later, but not until the same year is used for both a Census and an Electoral Register entry. By that time a large number of facts may have already been entered all using the same Fact for both Census and Electoral Register. That would then require the Electoral Register 'Census' facts to be converted to 'Residence' facts to be consistent with those entered in Census years and subsequently. Ouch!

Maybe Nick could comment on this topic.
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry
User avatar
NickWalker
Megastar
Posts: 2597
Joined: 02 Jan 2004 17:39
Family Historian: V7
Location: Lancashire, UK
Contact:

Re: Electoral Register v Census v 1939 Register

Post by NickWalker »

It is difficult really to know easily how to highlight this - I could write it all over the help file but in my experience most people don't seem to read the help and to be fair I'm not really someone who reads guides either! I should also add that I hate keeping the help pages updated - I find it tedious and time-consuming and don't enjoy it. I do enough of that kind of thing in my day job and I don't know want to spend my spare time doing more of it. I do the minimum I feel I can get away with. So I'm the first to admit that I could do a lot better at adding more to the help pages. But then as I said, I do find a lot of people don't read it anyway so that doesn't help my motivation :)

I'd like to imagine that the fact that the Census entry part of AS is being used would be a clue that the records to be created are census records.

There isn't a 'conflict' - it's just a validation check - pointing out that there is already a census fact for that year in case you've chosen the wrong year by mistake, but you can create as many as you like for the same year.

Some people might choose to create 1939 register entries as if they are census records using census facts (I certainly do), others may do the same for electoral registers as they do serve a similar purpose to a census from a family history viewpoint. I am also aware that some people use residence facts for censuses. Others may use custom facts for 1939 register or electoral registers, etc. Ancestral Sources caters for all these tastes - you can specify that a census fact, residence fact or custom fact can be used for each 'census type'.

Nick
Nick Walker
Ancestral Sources Developer

https://fhug.org.uk/kb/kb-article/ancestral-sources/
User avatar
tatewise
Megastar
Posts: 28341
Joined: 25 May 2010 11:00
Family Historian: V7
Location: Torbay, Devon, UK
Contact:

Re: Electoral Register v Census v 1939 Register

Post by tatewise »

It is not so much that users don't know what type of entry they are creating but that they won't know that all types use the same GEDCOM Fact, i.e. always Census or always Residence depending on the Tools > Options setting.
It only becomes apparent if they inspect the Census Templates (unlikely) or AS reports an existing fact for the same year as Colin found, which I mistakenly interpreted as more serious than just a validation warning.

Would it be feasible for an informative message to be given when the first Census or Electoral Register entry is chosen to say that both types are using the same GEDCOM Fact, i.e. Census (or Residence)? So distinguishing the two types becomes more difficult in Queries, Plugins, etc, and also may lead to validation messages.

However, that leads to further observations...

I have not seen the validation warning but is its wording ideal for the new situation of both Census and Electoral Register types of entry being created? i.e. Is it still worded as if only Census entries can be created and thus the validation warning is highly likely to indicate an error? Whereas now it is perfectly valid for the two types to exist in the same year.

For my Lookup Missing Census Facts plugin, the Census events are crucial. It checks the Date of the Census event against the actual date of the census, but it accepts a match even if only the Month & Year or just the Year agree. Also, a few censuses only have a year defined. So the plugin might match an Electoral Register entry is if it were a Census entry and omit that year from the missing census list. As far as the plugin is concerned it is just another Census event.

If I had realised that AS was 'overloading' the Census event in this way I would have said something earlier.
The Ancestral Sources Template ~ Electoral Register UK adjustments for AS V5 advise using the Residence fact and I assumed that would continue with AS V7. So there is a backwards compatibility problem for users who used that technique in AS V5 when they come to use AS V7 and assume it will continue to use the Residence fact only to discover sometime later that it is using the Census event instead.

Sorry, but I think there is more to this issue than just a validation warning.
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry
User avatar
NickWalker
Megastar
Posts: 2597
Joined: 02 Jan 2004 17:39
Family Historian: V7
Location: Lancashire, UK
Contact:

Re: Electoral Register v Census v 1939 Register

Post by NickWalker »

I think the wording for the validation message is OK but I will double check it works correctly in the circumstance where there is more than one census fact in the same year.

I did give some thought to whether to have it default to residence but decided that really a residence fact isn't necessarily the correct fact to use for this - we use a 'census fact' to show it is a census record rather than necessarily where they live and I understand that electoral register entries can be their business address rather than where they lived. There isn't an equivalent standard fact so I could imagine using a custom fact for that, but I didn't think it was the correct decision for AS to just assume they would use a custom fact.

It is also entirely possible to create other templates and 'census types' so there's nothing to stop people creating other types (similar to the 1939 register) that use census facts.

I can see the problem it would cause for your plugin and perhaps having the electoral registers default to residence would be 'safer'. This is why I was so disappointed that you didn't get involved in any of the AS v7 testing. I knew that once I'd released it you would start telling me all the things I'd done wrong! :D
Nick Walker
Ancestral Sources Developer

https://fhug.org.uk/kb/kb-article/ancestral-sources/
User avatar
tatewise
Megastar
Posts: 28341
Joined: 25 May 2010 11:00
Family Historian: V7
Location: Torbay, Devon, UK
Contact:

Re: Electoral Register v Census v 1939 Register

Post by tatewise »

I'm sorry I could not get involved in AS V7 testing but my 40 plugins have taken precedence and still need my attention.

I understand your choice and that users can configure AS in various ways. However, they may not be aware of all the options and the consequences of the choices until after a lot of data has been entered. Is there any way, like the initial message I suggested, that could make users aware that they need to opt into alternative options beforehand?
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry
User avatar
tatewise
Megastar
Posts: 28341
Joined: 25 May 2010 11:00
Family Historian: V7
Location: Torbay, Devon, UK
Contact:

Re: Electoral Register v Census v 1939 Register

Post by tatewise »

The more I mull this issue over the greater are my concerns.
It seems that the default settings offer no way to reliably distinguish a Census entry from an Electoral Register entry.
They all use the same standard Census Event with no unique field that can be tested to differentiate them.
There is a large body of published resources (and no doubt private user resources) that rely on the Census Event as representing a national Census. That includes Queries, Plugins, Diagrams, Reports, Property Box tabs, and KB advice.

All the FHUG KB Census Queries I have inspected use CENS[year=yyyy] to identify a national Census in a particular year.
There are Plugins that similarly only check the CENS year to recognise national Census entries.
The Diagram Text Scheme CENS[1+] looping index will lump all Census and Electoral Register entries together.
Diagram Box Conditions often only test CENS[year=yyyy] to enable Census Icons.
The standard FH Individual Census Report does not distinguish Census from Electoral Register entries.
Some FHUG KB Property Box custom Census tabs rely on CENS[year=yyyy] to list national Census years.
Much FHUG KB advice suggests using CENS[year=yyyy] to identify national Census years.

Some of the above could be updated to test specific national Census Dates instead of just the Year, but that is a significant amount of retrospective updating.
Others have no workaround such as the CENS[1+] looping index and Property Box tabs that only allow Data References so Dates cannot be tested.

Once a user discovers those issues they might like to convert their Electoral Register Census Events to some other Fact, but the Change Any Fact Tag plugin won't help much as it cannot differentiate them either.

When Electoral Register records were discussed years ago, my recollection is that the majority were Residence related and the few that were business related could be immediately changed after data capture.

I may be over-reacting, but I have a nasty feeling that this is a time bomb that will erupt after users have captured a significant number of Electoral Register entries and discover that Queries & Plugins work incorrectly, Diagram Icons popup unexpectedly, Property Box tabs are misleading, etc, etc...

My feeling is that the two types of records should use different Facts by default. The user has the option to change that but that to me is the better default. The only other ideas I have are to use the INDI.CENS.TYPE Descriptor subfield or the generic Source Type field to differentiate the two types of record. However, those ideas do nothing to resolve any of the issues with the existing body of resources.
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry
User avatar
AdrianBruce
Megastar
Posts: 2090
Joined: 09 Aug 2003 21:02
Family Historian: V7
Location: South Cheshire
Contact:

Re: Electoral Register v Census v 1939 Register

Post by AdrianBruce »

tatewise wrote: 21 Apr 2021 09:47 ... When Electoral Register records were discussed years ago, my recollection is that the majority were Residence related and the few that were business related could be immediately changed after data capture. ...
Not sure how relevant this is, but the further back you go - into the 1800s, I guess - the more likely it is that an entry in a Poll Book / Electoral Register, etc, refers to someone having a vote in a constituency relating to their ownership of property there, and not residence there.

IIRC the chap who owned the land that my house is built on, had a vote in the South Cheshire constituency in the 1800s, but is shown as being resident in the Isle of Man. That was obvious, it has to be said.
Adrian
User avatar
tatewise
Megastar
Posts: 28341
Joined: 25 May 2010 11:00
Family Historian: V7
Location: Torbay, Devon, UK
Contact:

Re: Electoral Register v Census v 1939 Register

Post by tatewise »

That relates to the 1832-1917 Electoral Register template in AS and most likely to clash with Census decades 1841 to 1911.
Adrian, what Fact Type is applicable to those Electoral Register entries that are ownership of property and not residences?

Currently, both Electoral Register and Census data use the same Census Event, so even identifying which ones to change to another Fact Type is a challenge. The AS template options let you change the default Fact Type for the two data entry types but it may take users some time before they are aware of the issues and already have a significant number of indistinguishable Census Events that need changing.
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry
User avatar
AdrianBruce
Megastar
Posts: 2090
Joined: 09 Aug 2003 21:02
Family Historian: V7
Location: South Cheshire
Contact:

Re: Electoral Register v Census v 1939 Register

Post by AdrianBruce »

tatewise wrote: 21 Apr 2021 10:25 ... Adrian, what Fact Type is applicable to those Electoral Register entries that are ownership of property and not residences?

Currently, both Electoral Register and Census data use the same Census Event, so even identifying which ones to change to another Fact Type is a challenge. ...
Good question. My primary Fact Type for Electoral Registers is a custom fact type "Voter" - i.e. a fact to indicate that the person is a registered voter in the place, where the place is a constituency, not a settlement-type place. And I recognise that isn't a road that many people will want to go down. I'd then usually(?) squeeze a Residence Fact out of it because I think that such a Register will normally indicate where someone actually lives, as well as the qualifying property. I guess "Possessions" is another fact type that might be useful to describe the qualifying property - but that's possibly another thing again.

So in other words, yes, I see the problems - I just wanted to ensure that people didn't imagine that it was only (local, qualifying) residences and businesses that are involved. And how you deal with existing stuff is a b*******n.
Adrian
Post Reply