Clarify What GedcomX Is

[joeflint]

I have been lurking around this project for several weeks. I am confused as to what we are trying to do.
Are we trying to:

1. Create a genealogical data transfer model for existing applications and services? If so will the model be designed to the lowest common denominator or is it to include all the data the message sending system can create? What mechanisms will be in place to extend the model as needed?
2. Create a data model that future applications and service should use to build data stores?

I assumed GedcomX was about option 1. However all the discussion about n tier for example leads me to believe that part of the community are working on option 2

[ttwetmore]

A very good question.

My assumptions have been that GEDCOM-X is

1. A data model and accompanying file format for archiving and transporting genealogical data.
2. That the model will be a superset of the models used by current desktop and on-line systems, to the extent possible, to enable data sharing between any pair of systems. This is impossible in the current world, of course, so we anticipate a future in which current systems are enhanced to import and export GEDCOM-X files, with the hope that the enhanced systems will also extend their internal formats to take advantage of more parts of the GEDCOM-X data model than their current versions do. To fully share data between systems they obviously must all implement the full GEDCOM-X model, too much to expect in the near future.
3. That the model will encompass both the needs for recording evidence and the needs for recording conclusions, and the needs for connecting the conclusions to the evidence. The needs for recording evidence include objects for sources and the evidence found in the sources, allowing citations to be generated. The persona is one of the objects proposed in this model, an object to record the information found about a single person from a single item of evidence. The needs for recording conclusions include person objects that refer to the record-level objects (personas particularly and their sources) that provide the bases for the conclusions. The N-tier approach is one of the proposed ways to connect multi-layered conclusion level objects to record level objects.

A couple big questions are in here. For example, does it really make sense to try to create a superset model for the industry, if we believe that the industry will not embrace it, or that it would take 20 years to reach that point?. The AncestorSync product takes an alternate approach by providing custom software that simply translates data as well as possible from one format to another with as little data loss as is possible. This is a solution coming along now, and if it works well enough, will the major pressure for a new data model to solve the data sharing problem go away?

Another big question is whether or not the data model should be extended as far into the record/evidence area as the current GEDCOM-X model suggests. The majority of all desktop and on-line systems today are conclusion-based systems, so why should the new GEDCOM-X model include the record model? Some of us are very much in favor of this, since it makes it possible for genealogical systems to better support a richer research process, but if those features are esoteric to the greater number of users, it is hard to expect the industry to adopt them.

[EssyGreen]

My assumption is that it is simply a mechanism for transporting genealogical data from one application to another ... ie simply a new version of the existing GEnealogical Data Communication. I don't see how it is possible to dictate what applications should do in terms of data structure, format or storage mechanisms.

Furthermore, this will necessarily mean adopting a reasonably minimalist model in order to ensure that it is easily adopted by as many applications as possible (see #141)

[ttwetmore]

There is a distinction to be made between dictating what applications should do, and providing a more complex model (than GEDCOM) to enable future applications to share more types of data than it is possible to do today, and to provide more features (e.g., better support for recording evidence) than it is possible to do today.

I believe the over-arching goal of GEDCOM-X should be to enable the lossless sharing of data between genealogical desktop and on-line systems.

Today's systems allow import and export of their data in GEDCOM format. For most systems each import and export is a lossy operation because the inner model of the systems and the GEDCOM format do not line up perfectly. This problem is compounded by the fact that different systems 1) interpret GEDCOM differently and 2) add extensions to GEDCOM that other systems to not recognize. I believe it is GEDCOM-X's role to solve this data sharing problem.

I agree that getting systems to support the GEDCOM-X model does involve some level of dictating what an application should do, but do you see any other way to reach the goal of better data sharing? Doesn't deciding that GEDCOM-X will take a minimalist approach imply that the data sharing problem will not be solved?

How important do others see the data-sharing problem?

[EssyGreen]

I've really already said my bit on this in #141 but basically I don't think you will ever get round the "lossy" problem - although we can/should get round the ambiguity issues of GEDCOM 5.

[ttwetmore]

I missed 141. Solving the ambiguity problem can be done by firming up the specifications of GEDCOM.

Giving up on the lossy problem makes efforts like GEDCOM-X and B-GEDCOM pointless, IMHO, though it does allow GEDCOM-X to be a simple tweak to GEDCOM.

I hope the GEDCOM-X masters will clarify this point.

If you believe the lossy problem is not get-aroundable, what do you believe the fundamental goals of GEDCOM-X should be?

[EssyGreen]

what do you believe the fundamental goals of GEDCOM-X should be?

See #141 :)

[ttwetmore]

No thanks. If you can't say it in a single sentence, I'm not all that interested. Sorry.

[stoicflame]

The purpose of GEDCOM X has been stated as:

To define an open data model and an open serialization format for exchanging the components of the genealogical proof standard.

When we talk about "the components of the genealogical proof standard," we mean these:

• Search Reliable Sources
• Cite Each Source
• Analyze Sources, Information, and Evidence
• Resolve Conflicts
• Make a Soundly-Reasoned Conclusion

There's still some work needed to articulate the specifics of how GEDCOM X is designed to exchange those components. That work will be tracked at #141.

Re: the "lossy" problem: it is not a goal of GEDCOM X to define a model for all features of all current (and future) genealogical applications in an attempt to eliminate "lossiness." However, in order be relevant, GEDCOM X needs to identify and account for the features that are common to a majority of current genealogical applications. It also needs to provide a reasonable extension mechanism to account for features that are either (a) future or (b) limited to a minority of providers who still want to share losslessly amongst themselves.