I have my family tree in FTREE/2, which is out of support. It supports a mixture of secular English names and religious Hebrew name. I have a script to export the tree as GEDCOM, but when I tried importing it a few years ago, gramps couldn’t handle it. If there is a format that gramps can handle then I’d like to modify the script and try again. Ideally I’d like to do this on ArcaOS, but Linux is acceptable.
(Please include your Gramps version and ArcaOS,Linux with KDE)
Thanks.
Could you export a twig of your tree so we can see what the targets are?
A twig might have 1 person, their spouse, a child, and a grandparent. (No living people, please.) The person should have samples of typical data that was troublesome … including the Name in both secular English and religious Hebrew. At least some data should have Source citations to test that import.
What scripting language are you using?
And a search of FTREE/2 found a free (but not open source) FTree version 2.1 from 2016. Is that the software you mentioned? Their documentation indicates that they favor the Family Tree Maker dialect of GEDCOM.
This the the layout I currently use:
0 HEAD
1 SOUR FamilyTree for OS/2 - ExGedcom.ftx
2 VERS 1.2x
1 GEDC
2 VERS 5.5.1
2 FORM LINEAGE-LINKED
1 CHAR UTF-8
1 FILE metz.ged
1 DATE 13 Oct 2020
0 @I1@ INDI
1 NAME Seymour Jerome /Metz/
1 SEX M
1 BIRT
2 DATE 10 JUN 1944
2 PLAC Detroit
1 DEAT
2 DATE
2 PLAC
1 NAME שְׁמוּאֵל בֵּן ל״ביש
2 TYPE Hebrew Name
2 GIVN שְׁמוּאֵל
2 ROMN Shmuel ben Leibisch
3 TYPE Hebrew
3 GIVN Shmuel
The script is in OREXX, the precursor to ooRexx. The program is the Family Tree/2 that you found.
ask LLM/AI to help you with your script so that it export straight to Gramps xml, that way you do not have any limits and you can define yourself where you want different other types of information you might have.
Your alternative names goes straight to Gramps Alternative names, but you might have information that actually goes to a custom gedcom tag, that could be mapped to an existing Gramps xml attribute…
just a tip.
I am not certain that the format is the same, but Gramps supported a Ftree export since a while.
Does it mean it should be now deprecated ?
I’m concerned wit the other direction:nexporting frm Ftree/2 and importing to gramps.
Here is similar data entered into Gramps and exported as a .ged file. I combined the 2 individuals because it seemed better to have an example with multiple alternative names. And added nicknames (in both languages), a title and a suffix. (I do not read Hebrew. So I probably confused Given Name and Surname.)
The only odd parts are
_RUFNAME which is used to indicate when someone is “called by” other than the first name in the GIVN, and0 HEAD
1 SOUR Gramps
2 VERS 5.2.2
2 NAME Gramps
1 DATE 14 MAY 2026
2 TIME 11:17:36
1 SUBM @SUBM@
1 FILE /home/districtsupport/Documents/Ftree2_ex.ged
1 GEDC
2 VERS 5.5.1
2 FORM LINEAGE-LINKED
1 CHAR UTF-8
1 LANG English
0 @I0001@ INDI
1 NAME Seymour Jerome /Metz/ Jr.
2 TYPE birth
2 GIVN Seymour Jerome
2 SURN Metz
2 _RUFNAME Jerome
2 NSFX Jr.
2 NPFX Prof.
2 NICK Jerry
1 NAME שְׁמוּאֵל /בֵּן ל״ביש/
2 TYPE Hebrew Name
2 GIVN שְׁמוּאֵל
2 SURN בֵּן ל״ביש
2 NICK שמוליק
1 NAME Shmuel /ben Leibisch/
2 TYPE Roman
2 GIVN Shmuel
2 SPFX ben
2 SURN Leibisch
2 NICK Shmulik
1 SEX M
1 BIRT
2 DATE 10 JUN 1944
2 PLAC Detroit
1 DEAT
2 PLAC
0 TRLR
And how it re-imported:
Data lost in the round-trip through GEDCOM:
OriginAccording to the support notes online, there were significant problems with that format and he abandoned it. He even eliminated the ability to read the old format in the 2.x versions. (The intent was to force users to use a 1.x version to convert to GEDCOM before upgrading.)
FTree supposedly uses a GEDCOM dialect now instead. And so does Web Family Tree by Peter Moylan. (Although that domain has apparently expired since 2024. Only the Wayback Machine afterimages remains.)
Having an IMPORTER plug-in would be OK. But yes, I’d vote that the exportftree.py EXPORT should not be carried forward in any more of the 6.x branches. What do you think @Nick-Hall ?
A middle name is distinct from a middle name.
That structure doesn’t associate the Hebrew name with with its transliteration.
In Shmuel ben Leibisch, שְׁמוּאֵל (Shmuel) is the name and ל״ביש (Leibisch) is the patronimic; there is no surname. I don’t know how to handle patronimics and matronimics in gedcom.
I have not seen a GEDCOM annotation for Patronymic. Gramps stores this in the “Origin”.
From Perplexity:
In GEDCOM 5.x
No standard NAME subtag means “this is patronymic.”
Common practice is to place it in the surname portion for compatibility with other genealogy software.
Some programs add custom tags or internal metadata, but those are not portable standards.In GEDCOM X
Patronymic can be represented explicitly with a name-part qualifier.
That makes it much clearer semantically than legacy GEDCOM.For a legacy GEDCOM file, the safest approach is usually to enter the patronymic in the surname slot and add a note or custom tag if your software supports it.
So if your script can add a custom tag, you could transmit it that way. Gramps will preserve all unrecognized tags as Notes. So you could run a post-processing SuperTool script to set the origins.
Or you could rely on the ben/bat/bar to be a marker and use a SuperTool script to set the origin. (An example of a script to set surname Origin to “Patrilineal” or “Taken” (and Type to “Birth Name” or “Married Name”) based on contextual matches to the Father and Groom surnames is shown in the article: “Tree vivisection experiments with the Isotammi SuperTool”)
I took a peek at Metz.fam, and the structure is very different from gedcom:
TREEUSER BHebrew Name CSSN BAKA BYahrzeit MANSðJ AMetz BSeymour Jerome ICDetroit KSystems Programmer L4963 Oriskany Drive, Annandale, Virginia 22003-5141 D
˜ N Shmuel ben Leibisch r ù J AShulman BLee ICKozhangorodok, Polish White Russia HSan Bernardino, California E18 Elul, 5740
SSN xxx-xx-xxxx (lists birth as 3 Feb 1912)
Remarried 6/21/1964 D y G ¼ N Leah bath Chanah vZalman Dov =û 18 Elul 5740 J AMetz BLeonard IHSan Bernardino, California KJournalist EDivorced 1957 or 1958
Remarried 6/21/1964
Yahrtzeit 30 Nissan, 5752 D
GÈ N Leibisch ben Aharon 30 Nissan 5752 J AKatz BYetta ICGolog¢ry, Galician Poland (now Belarus) about 20 km East of Lvov Ehttp://data.jewishgen.org/jgen/wc.dll?jgproc~jgsys~jripllat
AGAD JRI Akt: 5 Signature Event: 204
Notes say 5/?/1949 but gravestone says 12/9/1949 Kislev 18, 5710 (Friday)
As an OS/2 system, that is probably using IBMPC / code page 437 for the character encoding. Which does not clipboard cleanly. So that sample doesn’t really clarify.
As long as I know what gramps can handle, I can modify the script to suit.
Yes, I use CP 437/850, and OS.2 Unicode support i very minimal.
You could append something unique (like {patronymic}) to the name TYPE. That would be easily searchable to use as a flag for a SuperTool script to set the Origin. It would not flag as an error during import either.
And you could either simultaneously clean out the flag within that script… or use the Types Cleanup addon utility to do verification later then cleanup the custom type.
1 NAME Shmuel /ben Leibisch/
2 TYPE Roman
2 GIVN Shmuel
2 SPFX ben
2 SURN Leibisch
1 NAME Shmuel /ben Leibisch/
2 TYPE Roman{patronymic}
2 GIVN Shmuel
2 SPFX ben
2 SURN Leibisch
If it is an obsolete format, then we should probably remove it or move it to the third-party addons repository.
0014222: ‘Web Family Tree’ ‘.wft’ format is deprecated. Move plug-in from core to addons