Test file of extensions

[tychonievich]

The specific extension-related content is described in HEAD.NOTE. Note that not all of these cases are explicitly identified in the spec; notably, it includes

• extension tags aliasing standard tags
• relocated standard enumeration values, to parallel relocated standard structures
• relocated standard structure as a record, possibly at odds with this part from the spec (emphasis added) "These are called relocated standard structures and can only appear with superstructures that are not documented as a superstructure of that structure type in this specification"

Related to #87 -- if we like this file, we can use it to fill out maximal70.ged

[tychonievich]

On the aliasing tags (i.e. using _USER to mean SUBM, _JOUR to mean COMP, and so on): I think these violate the spirit of the extension versus standard section of the specification, but I couldn't find any specific prohibition against them there. If we decide they are in violation of that, we should update that section to be explicit about that and remove them from this example file. One reason not to prohibit them, and thus to leave them in the example file, is the expectation that version 7.1 might make standard something that some 7.0 files include using extensions; if we fully ban aliasing that will violate the "every 7.0 file is a valid 7.1 file" principle.

[albertemmerich]

1. Yes, aliasing tags violate the spirit of extension versus standard section.
2. If we think the actual standard version does not explicitly prohibit aliasing tags, we cannot prohibit them in 7.x: In that case files which we think to be valid under actual spec 7.0 would be invalid after update.
3. There are very popular violations of this kind in the wild. See FTM and their _MREL and _FREL describing the kind of relation of a child to its father and mother, used in FAM records as subtags of tag CHIL. Standard requires to use PEDI in INDI record, and not any tag in FAM record. At import FTM converts adopted children as presented by PEDI in INDI records to biological children creating wrong _FREL and _MREL payloads. So this aliasing tags are cause of severe data transfer problems. Seeing this it would be very desirable to forbid these tags. However I think this is too late for 7.x.
4. I see another problem what "aliasing" really means - is it allowed to use an extension if you have a very small deviation in the definition? Saying the extension tag is very similar but not identical to a standard tag? If we want to find a definition what is forbidden and what is allowed I think we have to do some work - it isn't easy at all! Summary: I think we are not ready to ban aliasing tags.

[tychonievich]

To do:

• add to maximal.ged (changing identifiers to avoid collisions)
• link from tools page

[dthaler]

Agreed in GEDCOM Steering Committee meeting April 11, 2024. Separate issues will be filed for the todos:

1. add to maximal.ged (changing identifiers to avoid collisions)
2. link from tools page