1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

  2. Only registered members can see all the forums - if you've received an invitation to join (it'll be on your My Summary page) please register NOW!

  3. If you're looking for the LostCousins site please click the logo in the top left corner - these forums are for existing LostCousins members only.
  4. This is the LostCousins Forum. If you were looking for the LostCousins website simply click the logo at the top left.
  5. It's easier than ever before to check your entries from the 1881 Census - more details here
Cookies Not a member of the forum? Are you at the right site? Is this the right site? New feature!

FTAnalyzer and TMG

Discussion in 'Family Tree Analyzer' started by jbuchanangb, Oct 20, 2013.

  1. jbuchanangb

    jbuchanangb Member

    Hello

    Just thought I would give you the benefit of my experiments.

    I have exported my GEDCOM file and the analyzer tells me that I have 8376 individuals in 2308 families.

    It found 8985 census facts, of which 2 were unusable, and 1114 residence facts of which 623 were treated as census facts.

    I investigated the 2 unusable census facts and found that they were both Iowa State census facts from 1895.

    It also reported some other apparent data errors, including one case where a person had been baptised as an adult (described as "of riper years"). She married in 1747, was baptised in 1749, and had a daughter in 1757. There is a GEDCOM tag CHR used for christenings, and a different one BAPT used for baptisms. This caused 5 data errors to be reported.

    I have 2769 entries in the Lost Cousins system, and, needless to say, there were no Lost Cousins facts in my GEDCOM file.

    By looking in the forum I discovered the required GEDCOM format for a Lost Cousins fact, and created a new Tag Type in TMG to generate such a thing.

    I discovered that I had made two rather fundamental errors when entering census dates into my TMG database. Utterly inexcusable laziness on my part, partly exposed by FTA. I had entered the date of the 1901 census as April 1901, and the date of the 1851 census as April 1851. FTA announced that it had found over 100 suspect 1901 census entries, but tolerated them. It did not announce any suspect 1851 entries. The evidence, from the Colour Census Report, is that it disregarded all the dodgy 1901 and 1851 census entries.

    I decided that I would try to get maximum GREEN boxes for direct ancestors, mine and my wife's, (270 individuals) by correcting entries in my TMG database and inserting "Lost Cousin" tags. To achieve this I changed the root person to my daughter and used the Colour Census Report to view only Direct Ancestors, and sort in ascending Ahnentafel order.
    I went through and corrected 1901 and 1851 census dates, which helped a lot. I added 60 Lost Cousins tags. It may take me a while to enter the other 2709. It seems that a Lost Cousin tag is needed per individual, doesn't work as a family tag.

    I found that my practice of creating one family census entry in the TMG database with parents as principals and children or other household members as "witnesses" meant that the children/others did not have a CENS tag in the exported GEDCOM. In most cases I corrected this by creating a RESI tag with the census date, to make use of the FTA feature to treat these as census facts.

    I have noticed that FTA is reporting at least one 1841 census entry as detected when the person concerned does not have such a census entry.

    It is a bit annoying that the Colour Census Report reports RED for individuals for whom no Death date is available. For example I have an ancestor who was aged 65 in 1841, and is reported missing from all subsequent censuses up to 1911.

    I have tested the click to search Find My Past, and it works. I even found a missing entry for one George Smith in 1871. (Wife's great grandfather.)

    I have tried the "Show Missing from Census" report, and, even with "Direct Ancestors" selected, it gives me a bafflingly long list of 120 people, most of whom I am pretty sure couldn't have survived until 1881.

    Another strange thing I noticed, when I used the "1881 Census Report" from the "Lost Cousins" tab. The report is headed "1881 England & Wales Census Records on file already entered into Lost Cousins Web site". It lists 1293 Records/387 Families. I can't figure out how FTA can have this vast list, when I have only entered 60 Lost Cousins tags into my GEDCOM. Should the heading read "NOT already entered"?
    When I double click an entry on this list, I get an error message "Sorry non UK census searching of Find My Past isn't supported in this version of FTAnalyzer". I wasn't expecting that.

    The statistics report 24 lone individuals. I would quite like to find these people.

    If anyone would like to comment on my feedback, I would welcome it.
     
    jbuchanangb, Oct 20, 2013
    #1
  2. Alexander Bisset

    Alexander Bisset Administrator Staff Member

    Many thanks for the detailed feedback its always good to hear from people how they used the program. Niggles and bugs are usually found by people like yourself supplying this sort of info and this leads to fixes and new features. So thanks again for the feedback.

    BAPT vs CHR tags. I've not treated either as must be a child. The only such tag is treated as must be an adult and that is the CHRA tag which is the GEDCOM tag for Adult Christening. I'm not aware of what the official difference between a baptism and a christening is? The reason for possible data errors being reported is probably that you have no birth fact for that person and the option treat baptism/christening events is turned on thus it treats the baptism/christening as a proxy fact for the birth. Since you use TMG which is 100% GEDCOM compliant perhaps you could change that fact to an Adult Christening? Or enter you own proxy birth fact eg: she marries in 1747 so must have been say 16+ at the time which would give a birth fact of BEF 1731.

    I'll need to investigate the colour census report discarding "dodgy" census dates, I suspect if the tolerate option is on then it shouldn't so it's a problem if it is.

    Yes Lost Cousins tag is per individual I've not looked at what would be needed to enter this per family. The issue is always do you assume that a family tag means everyone in that family alive on that census has all been entered? I've erred on the side of caution so far and only looked at individual facts for Lost Cousins tags.

    I've noted various discussions on the TMG forum about census "witnesses" not getting exported to the GEDCOM, I think someone was looking into an addon to cater for that.

    The 1841 census entry detected in error will be a residence fact for 1841. Part of the problem of having residence facts treated as census facts is that you may get spurious results. If you spend the effort (and only you can judge if its worthwhile) you can tidy up dates as you indicated you'd started to do then turn off the "tolerate slightly inaccurate dates" option this should then only match residence facts that have an actual census date. Although this theory needs testing I'm not sure if I tested that idea.

    The no death date available is what I refer to as a "Loose" death. In fact you do have death information available but you haven't entered it. eg: you know they were alive in 1841 aged 65 (probably rounded) giving a birth date of BET 1771-1776. Thus a death date should be BET 1841 AND 1886. This is what the program should report for that individual under the Loose Deaths tab. This would then limit the upper years to 1886. NB. I'm planning in a future version to allow an option to assume loose deaths have been applied so the red would auto disappear for the 1891-1901 census as the program would have worked out they would likely be dead by that point.

    The missing from Census and Lost Cousins reports have undergone an overhaul for v3.0.0.0 and should be more accurate. I found a few bugs that sounds like what you are describing. In particular the headings are sometimes misleading in v2.2.1.1. These should be fixed in the new version hopefully out today.

    The double click on the Lost Cousins list saying it can't look up FindMyPast needs fixed. Really a double click on the Lost Cousins report should open a facts list rather than search the web. Thanks for the bug report on that one. I'll try to sneak that fix into v3.0.0.0.

    The lone individuals can be found on the families tab and sort by family ID they will have a family ID of "Unrelated".

    Thanks again for the feedback. As I say I'm hoping to get the new v3.0.0.0 with all the new mapping features out today. I'm just awaiting feedback from one of the beta testers on a bounding box geolocation issue.
     
    Alexander Bisset, Oct 21, 2013
    #2
  3. Alexander Bisset

    Alexander Bisset Administrator Staff Member

    Yeah that was quite easy fix. Implemented for v3.0.0.0.
     
    Alexander Bisset, Oct 21, 2013
    #3
  4. Tim

    Tim Megastar and Moderator Staff Member

    I'm certainly looking forward to the new release. :)

    And great feedback JBuchananGB.
     
    Tim, Oct 21, 2013
    #4
  5. Liberty

    Liberty LostCousins Megastar

    Re baptism/ christening. I suspect that the difference is largely theological or at least church admin. (I may be completely wrong, mind you!!)
    I rather think that christening is the giving a name (N or M, for those who ever looked at the catechism) and baptism refers to entry into the church. For denominations such as Anglican, Catholic, etc. that don't generally go in for adult baptism, the distinction is one to concern the clergy rather than the family historian. I think we can use the terms interchangeably without serious consequences.:)
     
    Liberty, Oct 21, 2013
    #5
    • Agree Agree x 1
  6. Alexander Bisset

    Alexander Bisset Administrator Staff Member

    I think the you are probably right Liberty and that a Christening is the giving of a name a Christian name. However it can clearly happen as an adult too, hence adult christenings. In fact its a matter of annoyance for me that people often mistakenly ask for a christian name when they actually mean forename. Indeed if I fill out a form and it asks for a christian name I leave it blank or write not applicable.
     
    Alexander Bisset, Oct 21, 2013
    #6
  7. jbuchanangb

    jbuchanangb Member

    I have done a bit more work on cleaning up my database. Changing dates of census entries mainly so that FTA recognises both 1901 and 1851 census entries. I am definitely not getting warnings about 1851 census dates being wrong, when I import the GEDCOM into FTA. I am getting over 200 for the 1901, and although errors are reported the entries seem to be ignored, at least with respect to the Colour Census report.

    I have looked very hard at some of the records for which FTA is claiming to have detected an 1841 census entry, and several of them definitely do not have such an entry. Consider the following:

    and
    There is no 1841 entry in there, but FTA shows a YELLOW box for that census column. (Notice I have corrected the 1851 census date!)

    I could probably offer up some others.

    I hope this feedback is useful.
     
    jbuchanangb, Oct 21, 2013
    #7
  8. Alexander Bisset

    Alexander Bisset Administrator Staff Member

    In order for me to reproduce the error could you extract the couple of individuals listed above into a GEDCOM file then load that file and confirm it still shows the problem. Then create an issue at the programs website attaching the GEDCOM. This will add it to my to do list and keeps all the issues in one place. It's also handy for future reference to be able to track back an issues list rather than trying to find the report in an email or forum post or private forum message etc.

    It may be that the problem is already fixed hence the request once you create the GEDCOM that you double check that the problem still exists when the file only contains a couple of individuals. If you can see the problem in v2.2.1.1 then I can check if it still exists in v3.0.0.0 (almost ready to launch tonight just awaiting a go/no-go from launch control (aka my beta testers).

    Once I can reproduce the problem I should be able to fix it. I suspect it is an issue with invalid country names in a Lost Cousins census year a bug I squashed in v3.0.0.0.
     
    Alexander Bisset, Oct 21, 2013
    #8
  9. Alexander Bisset

    Alexander Bisset Administrator Staff Member

    Just pasting the raw GEDCOM there into a file I'm getting the following...

    forum yellow colour issue.png

    So as you can see it looks ok. Thus either I've already fixed the bug in v3.0.0.0 or the GEDCOM snippet above didn't contain enough info to demonstrate the bug.
     
    Alexander Bisset, Oct 21, 2013
    #9
  10. jbuchanangb

    jbuchanangb Member

    My version of FTA has updated to 3.0.0.2. When I extract single individuals the FTA gives the right result. I have managed to extract a GEDCOM with only 12 individuals in it, and I am seeing the problem with the 1841 census. I see it with several records. It seems to me that if the parents are in the 1841 census, then the child is considered to be also.

    I will post this as an issue on the web site
     
    jbuchanangb, Oct 21, 2013
    #10

Share This Page