FTAnalyzer questions.

[Bryman]

What is the difference between the ADDR and PLAC tags in a gedcom file?

Any file produced by GenoPro seems to put a company name in the ADDR record and location information in the PLAC record. Is that correct? Is that as expected by FTA?

One instance, where I had the company shown as "T/A Nexus" with no location information, resulted in geocoding for somewhere in Japan!!!
Adding the missing location data produced the required geocoding in UK.

 

[Alexander Bisset]

According to the GEDCOM 5.5.1 standard Address is structured whereas place is typically a single line of text. Address structure is typically used for things like source repository address, submitter address, address of company writing GEDCOM software etc. However it is perfectly valid to be included in a fact.

ADDRESS_STRUCTURE:=
n ADDR <ADDRESS_LINE> {1:1} p.41
+1 CONT <ADDRESS_LINE> {0:3} p.41
+1 ADR1 <ADDRESS_LINE1> {0:1} p.41
+1 ADR2 <ADDRESS_LINE2> {0:1} p.41
+1 ADR3 <ADDRESS_LINE3> {0:1} p.41
+1 CITY <ADDRESS_CITY> {0:1} p.41
+1 STAE <ADDRESS_STATE> {0:1} p.42
+1 POST <ADDRESS_POSTAL_CODE> {0:1} p.41
+1 CTRY <ADDRESS_COUNTRY> {0:1} p.41
n PHON <PHONE_NUMBER> {0:3} p.57
n EMAIL <ADDRESS_EMAIL> {0:3} p.41
n FAX <ADDRESS_FAX> {0:3} p.41
n WWW <ADDRESS_WEB_PAGE> {0:3} p.42

PLACE_STRUCTURE:=
n PLAC <PLACE_NAME> {1:1} p.58
+1 FORM <PLACE_HIERARCHY> {0:1} p.58
+1 FONE <PLACE_PHONETIC_VARIATION> {0:M} p.59
+2 TYPE <PHONETIC_TYPE> {1:1} p.57
+1 ROMN <PLACE_ROMANIZED_VARIATION> {0:M} p.59
+2 TYPE <ROMANIZED_TYPE> {1:1} p.61
+1 MAP {0:1}
+2 LATI <PLACE_LATITUDE> {1:1} p.58
+2 LONG <PLACE_LONGITUDE> {1:1} p.58
+1 <<NOTE_STRUCTURE>> {0:M} p.37

As you can see Place is typically just a line of text whereas address can have far more structured details. Place also can optionally have a MAP tag to record Latitude/Longitude. I've not properly handled ADDR tags yet as they are quite rare to find in events and often where they do exist are simply structured versions of PLACe tag. However I really should handle them and use them where a place tag is absent but an address tag is present.

 

[Alexander Bisset]

Edit: Correction I have handled ADDR tags (I must have added it a while ago and forgotten). Can you provide a GEDCOM snippet of the whole fact where the ADDR record was not loaded correctly. I can then check why its not handling the code as expected.

 

[Bryman]

According to the GEDCOM 5.5.1 standard Address is structured whereas place is typically a single line of text. Address structure is typically used for things like source repository address, submitter address, address of company writing GEDCOM software etc. However it is perfectly valid to be included in a fact.

Place also can optionally have a MAP tag to record Latitude/Longitude.

Thank you for this information. It does not seem to be the way that I would have expected it to be and I suspect that the developers of GenoPro may also have ADDR and PLAC round the wrong way too. I have queried this on the GenoPro forum for bugs and it has been viewed by several people but I have not yet received an answer.

Can you provide a GEDCOM snippet of the whole fact where the ADDR record was not loaded correctly. I can then check why its not handling the code as expected.

I am not sure that FTA actually needs any modification. It may just be that data is being specified to FTA incorrectly. I will come back with more news/requests when I have heard the explanation/suggestions from others.

 

[Bryman]

The Genopro developers have accepted the need for a change and posted the following information in reply to my bug report. Hopefully, this has fixed my problem and there is no need for a change to FTA. A little difficult to download and test at this time due to other happenings but will post more information when I can get back.

"Having revisited the Gedcom 5.5 standard it is abundantly clear that the AGNC tag is the correct place for the company/employer/organisation content.
AGNC tag content (responsible agent) definition:

RESPONSIBLE_AGENCY: = {Size=1:120}
The organization, institution, corporation, person, or other entity that has authority or control interests in the associated context. For example, an employer of a person of an associated occupation, or a church that administered rites or events, or an organization responsible for creating and/or archiving records.
I will fix this shortly and post a revised skin.
Update: version 2014.04.05 has the fix."

 

[Alexander Bisset]

Thanks for the feedback.

 

[Bryman]

Updated report skin successfully installed and company names, vicars at weddings, etc, are now being reported under the new tag rather than being shown as invalid addresses.

 

[Bryman]

One retrograde step noticed though as the new skin does not produce any Lost Cousins records in the gedcom file!
Was that an unofficial modification to the report generator in GenoPro which has therefore been backed out?

 

[Alexander Bisset]

Not sure Bryman, I think it was Tim that got that fix included. Tim?

 

[Tim]

Yes, there's some code that you add to the skin, I must have sent it to you before?

 

[Bryman]

Yes, there's some code that you add to the skin, I must have sent it to you before?

Possibly. I will go back and check in the morning. I had assumed that all previous updates had been amalgamated. Doesn't the report generator look for all tags that are present rather than just ones that it knows about?

 

[Tim]

Possibly. I will go back and check in the morning. I had assumed that all previous updates had been amalgamated. Doesn't the report generator look for all tags that are present rather than just ones that it knows about?

I found this that I posted.

So it depends which way you did it. If you did it by adding code, then you'll need to add the code again. If you followed the new method, then it should have worked and you could ask Genone as it to why it might have stopped.

 

[Bryman]

I found this that I posted.

So it depends which way you did it. If you did it by adding code, then you'll need to add the code again. If you followed the new method, then it should have worked and you could ask Genone as it to why it might have stopped.

Thanks for the reference Tim. I didn't do either of your suggested methods - I just installed the skin (Export to Gedcom 20130505) with the modification
already included. That is why I was surprised to find that the LC records were no longer being exported.

If that was what your attachment supplied, how do we get the code changes incorporated into the base product so that this does not repeat with all future updates? I notice that I was recently using the skin dated 20130730 and that still produced the LC records so perhaps the latest change has removed the necessary code? How can I check?

 

[Tim]

You need to contact Genome.
Genopro have not updated their software for years. Genome has been updating the skins to fix any reporting errors like this.

This is the primary reason that I moved to a newer different FHS. The secondary reason was FTAnalyzer!

Genopro is the best for visualising your tree but falls way behind on the data side.

 

[AndyMick]

OK - I've joined the party at long last! I was tempted by a friend's interest in his family tree, and having put it into FTM2011 I thought to download FTA and see what happened. BRILLIANT! Lots of typoing misttks instantly show up for starters.

First dumb question - on the census form, it look me some minutes (yes, the brain is AWOL ATM) to realise that the white and grey backgrounds are just to highlight families (can't see it mentioned in the documentation) - but what does red signify, please?

I note a touch of the Rumsfelds - all the unknown unknowns showing up - and I thought I was looking at Howsams

Other questions will no doubt follow! I was going ask another - how to make my Gedcom compatible (I convert from Excel for my ONS), but I can look at the FTM-produced Gedcom for that, so I will investigate that when I get back from this diversion!

 

[Alexander Bisset]

The colours should be hinted at from tooltip popups. Red means a direct ancestor, bold means known to be alive on this census.

It would be useful to capture your initial impressions as a series of questions so that this can enhance the documentation. A fresh pair of eyes asking questions is always helpful.

I'm not quite sure what the "GEDCOM compatible" question means. The program should read any GEDCOM file produced by a GEDCOM v5.5 compliant program. If your ONS data is just in a spreadsheet then that's somewhat more tricky to create a GEDCOM file from, as Excel won't know about family relationships.

Assuming you are talking about a spreadsheet you'd need a bit of programming to generate a suitable GEDCOM format file from the spreadsheet, not that difficult but not that straightforward either. It depends how rusty your programming skills are. Basically a lot of loops to write each field in a separate line. Of course once you have it able to export to GEDCOM you could load the file into any family history program as well as FTAnalyzer.

 

[AndyMick]

Thanks Alexander. I must be more patient, or less nervous - I hadn't spotted the tooltips!

Yes, I produce Gedcom from my spreadsheet already thanks to VBA and VB6. So in fact it's a question of making the data I have on the censuses show in the Gedcom that fact in a FACT (I couldn't find a smiley for "groan") I'll pursue this when I've done some more work on this new family.

 

[AndyMick]

DOH
Just found the colour codes documented on the Home page - I'd been looking in documentation!

The family I'm looking at has considerable migration to Canada - you correctly deduce the likely residence at census time, but is there a way to suppress the ones who've emigrated? (I know this doesn't always apply - one of them came back to England from Canada, via the USA!) but it would be useful to get the Canadians off the Missing from 1881 England page so I can see the true missing ones.

Also, is there a setting for a cut-off for people who have long since died? (eg. >110 years old must have died)

 

[Alexander Bisset]

I can't quite recall how you'd exclude people not in UK, I suspect if there is an emigration fact then it would exclude them but I'd need to revisit the logic to remember. It may well just be if they don't have a census record it highlights them, you could find them on the Canadian census for that year for example.

The cut off for those that are too old should be implemented there is a selection box on the census form that implements an upper limit on age, this defaults to 90 years. This does of course depend on the user having actually entered dates for birth/death events. Unfortunately all too often users have no birth or death facts for an individual despite knowing something about them. So for instance if you know Joe Bloggs is the father of Mary Bloggs born 1850, then he must have been born BEF 1834 (assuming you count a default of 16 years old to be a parent - some people will use 14 or even 12). Its a technique that few people use which is surprising as it is extremely effective at giving you narrow date ranges to concentrate searching.

This technique is what the Births/Deaths tab is all about it works out where you already know information about the individual and suggests ways you can narrow down the birth or death dates to more accurate dates than typically just unknown. It is what I call "Loose Births" or "Loose Deaths", by "tightening" up the date range based on known facts you can focus your research better.

 

[Alexander Bisset]

Yes, I produce Gedcom from my spreadsheet already thanks to VBA and VB6. So in fact it's a question of making the data I have on the censuses show in the Gedcom that fact in a FACT (I couldn't find a smiley for "groan") I'll pursue this when I've done some more work on this new family.

On your VBA/VB6 (blast from past) census facts export routine if you export it as a CENS fact this is what FTA is expecting. For extra info if you include a PAGE tag as an extra line on the census fact you can put the census reference there. eg:

1 CENS any comment here
2 DATE 1881
2 PLAC some parish, some town, some county, some country
3 PAGE Class: RG11; Piece: 3580; Folio: 59; Page: 32; GSU roll: 1341857

similarly if you have entered them on Lost Cousins you can add a

1 LOST COUSINS
2 DATE 1881

Then FTAnalyzer will pickup both the census record and the fact that you've entered the individuals census record on Lost Cousins. To add a page tag to the census record the format of the PAGE tag needs to match one of these regex.

EW_CENSUS_PATTERN = "Class: RG(\\d{2,3}); Piece: (\\d{1,5}); Folio: (\\d{1,4}); Page: (\\d{1,3}); GSU";
EW_CENSUS_1841_PATTERN = "Class: HO107; Piece: (\\d{1,5}); Folio: (\\d{1,4}); Page: (\\d{1,3}); GSU";
EW_CENSUS_1841_PATTERN2 = "Class: HO107; Piece:? (\\d{1,5}); Book: (\\d{1,3});.*?Folio: (\\d{1,4}); Page: (\\d{1,3});";
EW_CENSUS_1911_PATTERN = "^RG14PN(\\d{1,6}) .*SN(\\d{1,3})$";
EW_CENSUS_1911_PATTERN2 = "Class: RG14; Piece: (\\d{1,6})$";
EW_CENSUS_1911_PATTERN3 = "Class: RG14; Piece: (\\d{1,6}); Schedule Number: (\\d{1,3})$";
EW_CENSUS_1911_PATTERN4 = "Class: RG14; Piece: (\\d{1,6}); Page: (\\d{1,3})$";
SCOT_CENSUS_PATTERN = "Parish: ([A-Za-z]+); ED: (\\d{1,3}); Page: (\\d{1,4}); Line: (\\d{1,2}); Roll: CSSCT";
SCOT_CENSUS_PATTERN2 = "(\\d{3}/\\d{2}) (\\d{3}/\\d{2}) (\\d{3,4})";

PS. The main database application at work is a VB6 app I wrote 15 years ago so nothing really that wrong with VB6 apart from fact support is totally dead for it now. It's why I do almost everything in c# or VBA these days.