Modern Software Experience

2012-08-15

GEDCOM X not vendor neutral

GEDCOM X admission

Ryan Heaton, the FamilySearch programmer leading the GEDCOM X project, has just posted Requirements on the GEDCOM X blog.
It practically starts with the following statement (his bold):

I think all of the requirements for the project can be summarized into a single statement: GEDCOM X must be able to accommodate FamilySearch's Platform API.

implications

He then goes on to point out a few implications (again in bolded text, but I'll spare you that). These are:

and he adds that it does not imply that it doesn't mean that the goals of GEDCOM X are limited to the FamilySearch platform, and adds GEDCOM X may define concepts that FamilySearch doesn't intend to initially include as a non-implication.

FamilySearch platform

The GEDCOM X blog post mentions FamilySearch platform without explaining what that phrase means. It is nothing more than a grandiose sounding phrase for FamilySearch Family Tree (FS FT), their decade-delayed social genealogy site, originally known as New FamilySearch.
So, when he writes that GEDCOM X's one requirement is that it has to support the FamilySearch platform API, he is telling us that GEDCOM X has to support the FamilySearch Family Tree API, originally known as the NFS API.

Notice that it does not state that the GEDCOM X API must be identical to the NFS API, but merely that GEDCOM X must support the NFS API. Although FamilySearch is infamous for sloppy specifications, it would be a mistake to casually assume that this as sloppy writing and dismiss it as such. Ryan Heaton is a Java programmer, programmers tend to be annoyingly precise in their choice of words about things like this, and he wrote this blog post for a reason.

The GEDCOM X site makes it clear that GEDCOM X is the new FamilySearch standard.

the new FamilySearch standard

The GEDCOM X site makes it clear that GEDCOM X is the new FamilySearch standard.
GEDCOM X is not just the new data model, GEDCOM is the new file format and API too. GEDCOM X replaces the previous standards. GEDCOM X is the new FamilySearch file format, replacing GEDCOM. GEDCOM X is the new FamilySearch Family Tree API, replacing the NFS API.

NFS API worries

It remains hard not to wonder whether everyone within FamilySearch agrees with that viewpoint. After all, GEDCOM X is merely a work in progress, while the NFS API is in actual use. The NFS API, whatever it shortcomings, actually works while GEDCOM X does not really exist yet.

FamilySearch's positioning of GEDCOM X as the new FamilySearch standard has caused concern about the future of the NFS API among the FamilySearch partners using that API.
The GEDCOM X project was revealed in December of 2012. FamilySearch confirmed that GEDCOM X is their new standard at RootsTech 2012 in February of 2012, and FamilySearch Family Tree (NFS) went public on 2012 July 1.
For years FamilySearch told vendors to support the NFS API, and then, around the time that NFS goes into limited public beta, but still almost half a year before NFS goes public, they announce that the NFS API will be replaced already? FamilySearch announced that the API will be replaced even before the world had a chance to use the applications their partners created?
What the FamilySearch!

The NFS API is not obsolete. No replacement is happening yet. As long as GEDCOM X is not ready yet, making GEDCOM X the new standard is just a plan. Still, that plan worries vendors who put a lot of resources into developing applications using the NFS API.

NFS API reassurances

The Requirements blog post states explicitly that, although GEDCOM X is the new standard, the NFS API will remain supported. It even makes a considerably stronger statement. It states that supporting the NFS API is not just a requirement for GEDCOM X, but is in fact the requirement for GEDCOM X.

That must sound very reassuring to vendors who created applications using the NFS API. Vendors were aware that GEDCOM X is not just a file format, but an API too, and were increasingly wondering where FamilySearch's push for GEDCOM X would leave their current applications.
Surely, the major reason for the Requirements blog post was to reassure those vendors that the applications they created for the NFS API are in no danger of being obsoleted any time soon.

FamilySearch is promising backward compatibility.

FamilySearch is promising backward compatibility. The history of GEDCOM is no showcase for backward compatibility, but even that reality check does not provide reason to worry. There simply is very little reason to worry about compatibility, while GEDCOM X is merely a project in progress.

What should worry all vendors, not just the FamilySearch partners, is that the stated requirement directly contradicts previous FamilySearch statements about the nature of the GEDCOM X project.

open or proprietary

FamilySearch claims that GEDCOM X is an open community standard for the industry. The GEDCOM X site presents GEDCOM X as the standard for the genealogy industry. However, even after the 2011 article GEDCOM X revealed the existence of the GEDCOM X project, complete with the GEDCOM X source code, and vendors began looking over that source code to figure out what it is, FamilySearch kept mum about it, and even kept the GEDCOM X site closed to outsiders for another two months. Only two months later, during RootsTech 2012, did FamilySearch itself finally start telling the industry about their GEDCOM X project.

Former FamilySearch CEO Jay Verkler gave considerable attention to GEDCOM X during his RootsTech 2012 opening keynote.
The title of that keynote was Inventing the Future, as a Community, and during that presentation, Verkler took several minutes to argue that open is better than proprietary.

Never mind that some of Verkler's statements about GEDCOM are misleading at best, that GEDCOM does support Kanji as well Romanisation of those Kanji, and that GEDCOM is extensible.
What matters is that he not only confirmed the existence of the GEDCOM X project, but tried to create some pretty strong expectations for it as well.
He called GEDCOM X a A New GEDCOM, and clearly stated that it is a file format, an API and repository model. The GEDCOM X site states that GEDCOM X is data model, file format, and API.

He made it clear that he expects other companies, such as Ancestry.com, to support FamilySearch's GEDCOM X standard; the menu items on the transfer drop-down menu in his mock-up demo of the GEDCOM X future are Ancestry.com, Archives.com, FindMyPast, Geni, MyHeritage and More....

The Requirements blog post unequivocally states that the FamilySearch platform API is the all-encompassing GEDCOM X requirement.

open community project?

That may sound good, but as the industry watched the GEDCOM X project unfold, it has become increasingly clear that FamilySearch uses words like open or community in an unusual, FamilySearch-specific way.

The GEDCOM X site home page promotes GEDCOM X as Open standards for genealogical data communications and publishes some documents and source code, but that doesn't change the fact that it is a FamilySearch project for a FamilySearch standard.

The GEDCOM X article revealed that GEDCOM X started as part of the internal Data Framework. That bit of project history naturally makes you wonder whether FamilySearch continues to synchronise their internal Data Framework project with the public GEDCOM X project. FamilySearch has not bothered to publicly address that issue yet, but perhaps the Requirements blog just did.

The Requirements blog post unequivocally states that the FamilySearch platform API is the all-encompassing GEDCOM X requirement. To me, that statement is nothing less than an outright admission that GEDCOM X isn't open community project, but a proprietary FamilySearch project.

FamilySearch first

The Requirements blog post is clear: the FamilySearch platform comes first. What it says there is that the NFS API is more important than other APIs, because FamilySearch Family Tree (NFS) is more important than other social genealogy sites.

FamilySearch only

The blog post does not merely say that supporting the NFS API is the major GEDCOM X requirement, it actually says that supporting the API for their own social genealogy site is the only requirement. The other social genealogy sites do not matter less, they don't matter at all.
Other social genealogy sites have been around for years, are bigger, and their data is less unreliable, but the API for FamilySearch's Family Tree is the only one that matters to FamilySearch's GEDCOM X project.

That is an admission, from none other than the GEDCOM X project leader, that GEDCOM X is what practically everybody already thought it is: a proprietary FamilySearch standard. FamilySearch keeps trying to paint GEDCOM X as an open community standard, but actions speak leader than words. The FamilySearch rhetoric only fits when you understand open to mean FamilySearch-dictated and community to mean FamilySearch friends, followers and fankids.

abandonment risk

The GEDCOM X FAQ states that GEDCOM X will not go stagnant like all the other standard effort. FamilySearch's chequered history makes it hard to believe that already. It would not be the first time they announced something only to abandon a few years later.
The Requirement blog post gives more reason for worry. That FamilySearch abandoned both PAF and GEDCOM when it started work on NFS and the NFS API is no coincidence. They decided that the old stuff was not needed anymore now that they had decided to create something new.
One day, FamilySearch will replace Family Tree (NFS) with something else. It is not unreasonable to suspect that FamilySearch will abandon Family Tree and GEDCOM X together, just as it abandoned PAF and GEDCOM together.

GEDCOM X isn't a community standard. GEDCOM X is a proprietary FamilySearch standard for FamilySearch products.

proprietary standard

GEDCOM X isn't a community standard. GEDCOM X is a proprietary FamilySearch standard for FamilySearch products. The GEDCOM X blog just confirmed that.

FamilySearch is serious when they present GEDCOM X as the New GEDCOM. They want to pull another GEDCOM on the industry; they would like to regain the position of controlling the industry standard.
FamilySearch would like other vendors to support their system by supporting their standard. That would make the world less challenging for FamilySearch, but competing companies simply aren't in the business of supporting FamilySearch systems.
I am sure other companies have the same dream, about them creating their standard for their own use, and all other players adapting to what they created, without having to do anything in return. Most companies have enough sense of reality to understand that this dream scenario is just a fantasy.

The GEDCOM X project is not ready yet, but most, perhaps all, of the big decisions were made by FamilySearch before FamilySearch went public with it. FamilySearch is not supporting the creation of a community standard, it is trying to dictate a FamilySearch standard to the community.

Many in the industry understood this already, and are not enthusiastic for yet another FamilySearch standard. That explains why large and leading companies like Ancestry.com and RootsMagic have not joined the GEDCOM X project, but the Family History Information Standards Organisation (FHISO) instead.

updates

2012-08-31: FamilySearch drops standard claim

FamilySearch abandons GEDCOM X push, see GEDCOM X: no industry standard.

links