Adding BIO_FATH/MOTH as well as existing FATH/MOTH as ROLE values

[tychonievich]

As people and GEDCOM drift away from recording only genetically linked pedigrees and trend more toward family history we need to define terms that differentiate relationships that are biological from social. Some people may identify a person as a mother/father that is socially true but not biological (their parent!) and at the same time not acknowledge a biological association with a specific individual. GEDCOM has already indicated that a father/mother is a parent, so I think GEDCOM needs to introduce a new term for a biologically linked individual for example “bio_father” rather than using a phrase!

Such as:
1 ASSO @x123@
2 ROLL BIO_FATH
2 ROLL BIO_MOTH
2 ROLL SURROGATE

These terms can be added as enum rather than as a phrase!

Originally posted by @Norwegian-Sardines in #336

[tychonievich]

Discussed in steering committee

• We agree that both biological and social parents exist (and may participate in the same event, ASSO, etc), so it might be nice to have ROLE values to distinguish between them.

• There are more than just two categories of parenthood (adopted, de-facto, step, in-law, guardian, etc), but the number of kinds and the lines between them vary by culture, making them hard to enumerate. Even "biological" is complicated for mothers where the provider of the egg and womb might be different people.

• If we add BIO_MOTH in 7.1 (as opposed to 8.0), we'd need to be careful to make it clear that MOTH has not changed meaning and that BIO_MOTH is just a more specific subtype of MOTH.

  • An alternative might be some kind of more specific substructure, perhaps something like

    2 ROLE MOTH
    3 PEDI BIRTH

• We're not sure how to balance the various ideas and constraints above, and don't know if there are applications that do something like this now. If anyone here has such an application and can provide their input, we'd value that.

[jl5000]

I don't know why we couldn't just deprecate FAM.HUSB/WIFE/CHIL (possibly SUBM too) and just let the association structures deal with all of it.

[Norwegian-Sardines]

I don't know why we couldn't just deprecate FAM.HUSB/WIFE/CHIL (possibly SUBM too) and just let the association structures deal with all of it.

We can only do this in v8.0 GEDCOM since this change would break v7!

[jl5000]

Deprecate doesn't mean remove, it just means discourage. It provides users the time to transition gradually.

[Norwegian-Sardines]

Deprecate doesn't mean remove, it just means discourage. It provides users the time to transition gradually.

I know what deprecate means!

But you can’t move to a completely different relationship system this way! Don’t get me wrong I would like to deprecate the family_record concept in v8 GEDCOM!

[jl5000]

I really don't appreciate the tone of your responses. This seems to be a regular occurrence so I'm going to refrain from any more comments because I've had just about enough.

[tychonievich]

@jl5000 Thanks for your comments. It is true that there is a lot of appeal in removing the complicated FAM record and the many issues it's complexity induces. We've been talking about removing/deprecating that since the first few weeks of v7.0.0's preliminary design process. When we reached out to the companies listed in the contributors section, consensus was that removing FAM was not something to do in 7.0.

But could we add a better alternative and deprecate the old one? Yes... as long as we can do that in a way that doesn't create interoperability challenges. Our current model is that a minor version (like 7.1) can add new expressivity, but not change how 7.0 expresses the things that it does express. That is a more strict interpretation of semantic versioning than some projects have, but it does make interoperability stronger. It also means that we're hesitant to introduce an alternative way of doing something and deprecate the old way in a minor version.

All of that said, I at least do want to replace FAM eventually, and would love to see any ideas about a replacement. Using ASSO is interesting (and appealing because it would be easier to implement because it's already part of the spec), but it does leave unclear to me where things like PEDI and NCHI and birth order of children would go. Do you have an example of what that would look like?

[mother10]

All of that said, I at least do want to replace FAM eventually, and would love to see any ideas about a replacement.

Well I thought #679 did just that. It has no more FAM records. The family relation is now in the TEMPLATE. And the ROLEs in there, indicate the relation. With that its even possible to have a multi person relation, like Polyamory without any problems.

[tychonievich]

Discussed in steering committee

There appear to be two broad proposals in the conversation above

• The original post asked for adding a new ROLE value, which could be done in 7.1.
  • We were uncertain how to proceed on this and asked for suggestions in an earlier comment; that uncertainty remains.
• Several follow-ups propose making an alternative to FAM, which could not replace FAM until 8.0.
  • As a complexity, there's a discussion about whether it could be added as an option with a deprecation of the old option in 7.1. While the specification itself does not specify the rules when deprecation is appropriate, there is current use of deprecation to fix omissions (like EXID without TYPE) and to deprecate all but one semantically-equivalent way of expressing the same idea (such as deprecate ADR1 which is redundant with ADDR's payload). There are also notes in the specification indicating things such as "The FAM record will be revised in a future version," which is separate from (but related to) deprecation.

We recognize that FAM and its replacement is a complicated and important topic, but prefer to avoid mixing topics in a single issue so recommend that such conversations be directed to other issues (if there's a concrete proposal like #679) or discussion (for a more open brainstorm or abstract ideas), leaving this issue to track the proposed additional ROLE values in v7.1.

[Norwegian-Sardines]

All of that said, I at least do want to replace FAM eventually, and would love to see any ideas about a replacement.

Well I thought #679 did just that. It has no more FAM records. The family relation is now in the TEMPLATE. And the ROLEs in there, indicate the relation. With that its even possible to have a multi person relation, like Polyamory without any problems.

In my opinion, all relationship values must be codified by the GEDCOM Standard itself, rather than as a set of variables defined by the transmission. This is to say that the definition of what the term used for a specific relationship type (father, mother, half-brother/sister, foster-child, adoptive-child, etc.) must be defined in the GEDCOM Standard so that all created transmissions can be merged together and/or compared using a standardize term set and definition for each term value used. This includes for any language, custom, local and/or time period.

Unless I'm misunderstand the design presented in #679 these terms and definitions would be defined by the creating program and transmitted to other programs and merged with transmissions or local uses that could potentially set different terms/definitions for similar use-cases!

[mother10]

First remark: Luther asked to not mix discussing FAM here, but where it belonged in #679, and I wanted to follow that.
Next thing:

Unless I'm misunderstand the design presented in #679 these terms and definitions would be defined by the creating program

Yes, totally misunderstood. That is totally not whats in that proposel.
But I am not going to discuss that here, as Luther asked us.