Publicising Lost Cousins

[Alexander Bisset]

One thought - although manually adding additional household members after the first is a pretty trivial task, I know there are quite a few members who, to save time, have only entered one or two members of a household. Is there any way of automatically pulling in the census data for the other members of the household who are in the user's tree, then updating their My Ancestors page automatically?

If they have tagged the person in their tree as having a census then yes it will add those people to Lost Cousins. If they don't have a census record for that person it won't assume they should have.

That said it would highlight that those persons were alive on the census and are missing census info. It just cannot assume that the others in the family will appear on that census and short of looking up a site such as FMP it wouldn't be able to tell. Then we are in the realms of updating users data for them.

 

[peter]

Then we are in the realms of updating users data for them.

Not necessarily - they may well have printouts, or record the census information in some other way. In my own tree I record the information in my own way, and I'm sure I'm not the only one.

 

[Alexander Bisset]

Not necessarily - they may well have printouts, or record the census information in some other way. In my own tree I record the information in my own way, and I'm sure I'm not the only one.

What I meant was that if the user hasn't recorded a census record in their tree then there is nothing I can do to guess if someone on a census is of interest to them. In your own case, if I recall correctly, you have a lot of census notes and have recorded the census info as a cut n paste from a source into the notes for a family. That's fine as that will almost certainly include a valid census reference even if you haven't recorded a census fact against each person.

My basic point is that the user determines for themselves how they want to record their data and I cannot tell them who it is important to record or not. Trying to second guess what a user considers important it simply running the risk of antagonising them.

Thus the approach I've taken has been - if you have recorded a census for someone and there is a census reference that is in a format that is required for Lost Cousins then that is a candidate for updating. If they have chosen to ignore people from a census it is not for me to tell them they are wrong to ignore those people.

I can do a report to highlight that it might be an error to have missed them but I leave it to the user to decide what they want to record and what they don't.

The downside of this user control is that if they haven't bothered to record any census references then nothing I can do will get the info required for Lost Cousins to work as the fundamental principal with Lost Cousins is that they must have an accurate census record to match with someone, and I'm fairly sure that that is never going to change as to change would cancel the 100% guaranteed match promise.

So we can be fairly blunt - if you don't have a census reference you won't get to enter info into Lost Cousins manually or automatically.

 

[peter]

I haven't got census references in my tree for any of my relatives. It has never been necessary for my own purposes (I have digital images of the census pages which include this information), and as I input the details to LostCousins from a printout of the relevant census page/transcript it has never been necessary for LostCousins purposes either. Like many people who have been researching for a long time I have paper files as well as digital folders.

Whilst there probably won't be anyone else who has done exactly what I've done, I'm sure there will be many who haven't got census references in their tree, especially amongst the 20k members who have yet to make their first entry. If they have to add census information to the record for every member of a household, rather than just the head of household (say), they might well conclude that it isn't worth the effort - the average household has between 4 and 5 people.

So I was hoping it might be possible to help out these people, as well as anyone who has entered just one person from each household.

 

[Alexander Bisset]

It really does depend on how they enter the data. If they have entered the census entries as a family fact and have a census reference then it should be identified for everyone in a family.

HOWEVER. There is only so much hand holding you can do and if people are actively ignoring entering data in their tree then there is absolutely nothing that can be done to drag them kicking and screaming into entering data into Lost Cousins. Its exactly the same for the people you referenced that have their data on paper printouts. There's nothing wrong with having data on paper but if their priority isn't to enter data into a computer then there's nothing that can be done to get them to enter data onto a website.

There is no magic bullet here. The solution I'm offering is for people who are active researchers and have recorded census reference information but haven't transcribed that info into Lost Cousins format.

For some people the ability to automate the process might just be the nudge they need to enter the reference into their tree, especially when with just a double click they can get to the record online. For some the process will highlight where they are missing data and will give them an incentive to find it. For others even if you were to offer free lifetime subscription to Ancestry and FindMyPast they aren't going to put their paper records into a computer. That's just the way it is.


Note make sure you don't confuse two separate things here.

• The data in a GEDCOM will show facts for individuals some of these may include census references and so even if the person hasn't recorded a census the reference is sufficient to work out the census they have been found on.
• The data on websites that have census info and shows other members of the household

There is no link between other data on the website about different individuals and and the data in someone's tree. FTAnalyzer will analyse the data in a GEDCOM file is doesn't then go off online looking up census records to see if other records exist. It can only do what the user instructs it to do and going off online to look up data is something the user needs to tell it to do. At present there is no way without the user first having logged in for the program to search any of the websites and even when it does that at present there is no way for it to tell that other records on the same page refer to other people in their tree. For that to happen the program would need the user's permission to access the internet - trawl through website links (which of course means the user has to have entered the census reference first) then find a record then compare it with other people in that family then guess if it's a match.

Then EVEN if it did ALL of that it would still then require the USER to enter the data into their own tree in a format THEY control. As I said before FTAnalyzer cannot and will not attempt to edit someone's data and add stuff to their tree. So even if it found other people in the household the user would need to decide to add data to their tree. Again it's a case of leading a horse to water but not being able to make it drink.

If the user doesn't enter the data I can't force them to do so. Sorry. I don't know how many other ways I can say that.

 

[Alexander Bisset]

One thing that occurred to me though is that various programs (of which I think GenoPro is one) does allow for facts to be recorded against a family (the head of household) and this gets recorded against the entire family. This does get recognised by FTAnalyzer at present so yes in this case they would just enter one census reference to get the whole household.

 

[PhilGee]

Alexander,

I agree with your earlier responses: there are too many possibilities for errors in "doing the work for the user" regarding missing family members. This is shown in your message above where "household" could refer to different entities:
1. a group of people consisting of parent(s) and children where the "head" is a parent; but not necessarily living together at the time of a census
2. a group of people living together in a house (as would appear on a census return) where the "head" is often unrelated to your ancestors.

Unfortunately, I am one of the "lazy" clan: having checked census data to support my research but abjectly failing to record this in a way that supports entry into LC. This recent development of FTA is an incentive to bring my records and "trees" up-to-date with supporting references.

Phil

 

[Alexander Bisset]

Unfortunately, I am one of the "lazy" clan: having checked census data to support my research but abjectly failing to record this in a way that supports entry into LC. This recent development of FTA is an incentive to bring my records and "trees" up-to-date with supporting references.

That's a big part of the reason for doing this feature. If the user realises that with a little effort on their part that will tidy up their tree anyway and add useful information they can potentially quickly contact new Lost Cousins. We've all probably looked at the page at some point and thought I really must get round to entering the data but your records just aren't easy to organise in a fashion that makes the data entry process quick so you put it off.

The beauty of the approach I've taken is that you can see from a FTAnalyzer report which records have a census reference and which don't and so which ones you need to tweak. Then simply by adding that reference the program will automatically give you a new Lost Cousins entry. My gut feeling is that this iterative approach works well. You see a report to fix what's missing, you fix it then the report shows it can be auto added, you rerun the report and see fewer missing and repeat. All the while you see progress being made which encourages you to do more.

 

[peter]

"household" could refer to different entities

In census terms (and LostCousins terms) it only has one meaning.

 

[Alexander Bisset]

Right I've released v7.3 just now so it's live and people are starting to use it. My gut feeling is that the feedback from users will give a good steer on what changes are needed to make the process as smooth as possible for most people.

Meantime I'll get working on adding the forms to the Mac version. Note the code for doing the posting etc is all shared it's only the look and feel stuff (buttons/menus/grid displays etc) that's different.

I'm assuming Peter that you'd prefer people enter everyone in a household and not just the head of household and that's the focus of your comments?

 

[peter]

I'm assuming Peter that you'd prefer people enter everyone in a household and not just the head of household and that's the focus of your comments?

It's preferable for a number of reasons: firstly because people who enter a single won't necessarily enter the head of the household (they might not be a blood relative). Secondly because it reduces the chance of matches being missed because of errors in the data, or where the transcripts at different sites differ. And last, but not least, because it will improve the member's chances of getting an invitation to join this forum!

 

[Tim]

So we're seeing some great early results from Alexanders newest relaease of FTAnalyzer which includes the upload to Lost Cousins.

Here's what Debs has had to say.

 

[Pauline]

Unfortunately, I am one of the "lazy" clan: having checked census data to support my research but abjectly failing to record this in a way that supports entry into LC.

I'm not sure this necessarily has anything to do with being lazy. I don't have either census details or references in my family history program but I consider this to be down to historic realities and choices not laziness.

Don't get me wrong, I think what you are doing is brilliant, Alexander, and if I was starting out on my research today, I'm sure I would be using it. However, I started researching back in the dark ages when keeping paper records was the only viable option. Moreover, finding folk in censuses was a whole different ball game back then - no online censuses and indexes etc, nor any 'online' at all - so the only options were a trip to London or to the appropriate local record office and lots of trawling through films, often with limited success.

When I was able to go digital, for various reasons my initial focus was on building a website, and with early family tree programs having pretty limited capabilities, it just never became my primary record. I guess I could remedy that now but it would be a mammoth task, and I'm not sure I would reap sufficient benefit from doing it.

Meanwhile, I have already entered all possible known census entries into Lost Cousins, and now the issue for me is finding new relatives in the census - entering them is the easy bit!

 

[Alexander Bisset]

Oh absolutely Pauline if your records aren’t in a digital format as I said to Peter earlier there is absolutely nothing I can to to get the data into Lost Cousins. So yes for long time researchers such as yourself having a paper based system isn’t going to see any benefits from this update. With one possible exception...

That’s been just over 24 hours since the new version went live and already over 5000 new entries have been added. So the chances of you getting a match is increasing with each new entry added.

Indeed it may be worthwhile when others who are in your shoes come to use Lost Cousins next that you click search to refresh your matches just in case.

 

[Bob Spiers]

Alexander, advice please. Using FTA 7.3.0.0. (beta 16) (and other betas prior) - the opening upload of an Ancestry sourced Gedcom gives the following message:

Invalid GEDCOM.Line 121833: 'Othe' Error was: No space found in line: 'Othe'

At which point the upload stops and produces an error report and asks if I wish to view YES/NO. Clicking NO the upload completes with sundry minor error reports(mostly error parsing date ones) but then finishes and I can use FTA as required.

Clicking YES produces the Report in the form of the complete Gedcom (in read only form) and I can then scroll to line 121833 (which is blank) and see the actual reference is to line 121834 which shows (in bold if that is pertinent) Othe (and nothing else)

It is at a point in the Gedcom concerning an Irish ancestor MICHAEL FLYNN so line 121832 (a note made about him appearing to have two spouses) reads as : 1 NOTE There are two spouses for Michael Flynn..."...121833 blank: ...121834 Othe: ...12835 2 CONC r research shows Michael Flynn's ...etc

I am not that well acquainted with Gedcom format so have no idea where the 'Othe' comes from or whether to ignore or correct. I am aware I could open the Gedcom in Excel and just delete the 'Othe' and leave blank (or correct or complete the word). But that begs the question when I next upload an updated Gedcom from Ancestry (my standard practice about once a month) the line will still show as Othe and be picked up by FTA. I do not want to import a manually amended Gedcom into Ancestry, so it will likely occur again (and again).

I suppose the obvious solution would be to modify the Ancestry Page, but cannot even find where the note exists or the (part) word Othe? So what would you suggest I do please?

 

[Pauline]

Indeed it may be worthwhile when others who are in your shoes come to use Lost Cousins next that you click search to refresh your matches just in case.

Yup - been there, done that ...

It's something I have tended to do on a fairly regular basis anyway but it seemed like a good idea to up the frequency this last week. No new matches as yet, though.

 

[Bob Spiers]

Alexander Update: I have discovered the NOTE it is part of a Comment I made on Michael Flynn's page; "There are two spouses for Michael Flynn...etc..etc (completing the sentence) For whatever reason I then left about 4 lines before continuing: ..."Other research shows Michael Flynn's marriage...etc..etc (completing the sentence)

So playing detective it appears that the Gedcom records the first sentence, seems to dislike the extended blank lines and then records the second sentence starting: Othe (missing off the 'r') before continuing with 2 CONC r research shows....etc. (if you haven't already noted it, the missing 'r' is shown before the sentence begins).

I have now made both sentences into a flowing paragraph, no line gaps and so, fingers crossed, a new Gedcom should not contain the invalid 'Othe' line...hopefully! The proof of the pudding is to export a new Gedcom and upload it to FTA.

 

[Alexander Bisset]

Hi Bob,

It sounds like it's a classic extra carriage return error that gets added by Ancestry occasionally when you load data across. The check is simply to edit the text and replace any new lines with a space and/or a press enter to force it to be a "proper" line end rather than the spurious line ends that ancestry creates.

The new report you refer to which is optional to view allows you to view error lines in your GEDCOM. The bold is probably another spurious display character as GEDCOM is meant to be a plain text format.

 

[Bob Spiers]

Hi Bob,

It sounds like it's a classic extra carriage return error that gets added by Ancestry occasionally when you load data across. The check is simply to edit the text and replace any new lines with a space and/or a press enter to force it to be a "proper" line end rather than the spurious line ends that ancestry creates.

The new report you refer to which is optional to view allows you to view error lines in your GEDCOM. The bold is probably another spurious display character as GEDCOM is meant to be a plain text format.

Alexander,
Duly noted and at the end of the first sentence I pressed 'Enter' which, as you say, forces a proper line end. You may be pleased to know the newly imported Gedcom now opens without the invalid line. Thanks for your comments.

 

[Bryman]

and with early family tree programs having pretty limited capabilities, it just never became my primary record. I guess I could remedy that now but it would be a mammoth task, and I'm not sure I would reap sufficient benefit from doing it.

I do not wish to argue with you. The way that you work is your choice and not for me or anyone else to say that you are wrong. However, may I suggest that you could get some considerable benefit from progressively entering your paper held records into one of the fairly recent family tree programs. One major advantage that you would have over a new researcher is that you already have the information needed in your paper records.

Many years ago, I started with mainly paper records but now no longer refer to them since I have stored the information within my GenoPro system. That package may not be appropriate for you but there must be something which you would find to be more suitable than the previously investigated software with limited capabilities. I find that I now have a much better overview and understanding of my tree contents without the paper records. Off-line back-up is an added protection and does avoid the need for extra filing cabinet space.

Starting with just your direct ancestors may give you an alternative view of your tree contents and enable you to make use of utilities such as FTA to check that each family gets expanded to include all appropriate information. It may take some considerable time and effort to complete but would not need to be performed all at once.