Subject: GEDCOM and Formal Standards Organizations

From: Bill Harten

Reply-To: Genealogical Data Communications Specs

Date: Wed, 24 Jan 1996 11:53:52 -0700

This information will become part of the Frequently Asked Questions information available by anonymous ftp at gedcom.org.

People still ask: why wasn’t GEDCOM developed through a formal standards organization?

GEDCOM was first proposed at the National Genealogical Society(NGS) Conference held in Salt Lake City in July 1985. Objectives and principles were outlined in a first draft of the standard (1.0). Discussion at that meeting centered around how to organize to accomplish it, although no conclusions were reached.

I contacted officials of ANSI (American National Standards Institute) and NISO (National Information Standards Organization–Z39) in Washington D.C., as well as participants in ISO (International Standards Organization) standards, to learn what it would take to establish GEDCOM through their formal standards processes. I learned that these organizations require a minimum level of participation from a cross-section of industry representatives. The cost of their participation included fees, travel, and time.

I also visited with the chair of the Computer Interest Group of the NGS in Washington D.C., and invited them to co-sponsor the GEDCOM effort. This proposal was declined, partly because it was closely associated with PAF, and NGS could not endorse products.

We hosted the first GEDCOM developers conference in Salt Lake City in August 1987. About 30 invited guests attended, representing several different genealogical software products, user interests, and religious affiliations. Release 2.0 was presented. The Family History department then proposed establishing GEDCOM through the formal standards processes sponsored by ANSI or NISO (U.S.), and eventually ISO. We expressed our readiness to follow the formal approach, and asked who of the other developers could make the necessary commitment of time and money. At the time, all of these were very small organizations. None were prepared to participate in the formal process.

As an alternative, we proposed following the approach used by the U.S. Library of Congress in establishing the U.S. MARC format (MAchine Readable Catalog) for bibliographic information, now the dominant standard worldwide. Their approach was to define and implement a standard on their own, and publish it for others to follow if it was in their interest to do so. MARC began deliberately as a de-facto standard.

Similarly, the FHD offered to develop and publish the standard to meet our own needs, and to the extent possible, the needs of other independent developers and users. We agreed to add tags and structures where necessary, even if our own programs did not require them. We agreed that our FamilySearch programs would follow the standard for exporting data from our databases.

After considerable discussion, the GEDCOM developers raised their hands in unanimous support of this proposal, with a “wait and see” commitment to follow it if they liked it, and do something else if they didn’t. Thus GEDCOM was born as a deliberate, de facto standard, to be followed only by those who felt it was in their best interest to do so. The agreement remains to this day. We have been pleased with the voluntary results.

The FHD has not pressured developers to follow GEDCOM. The user community, on the other hand, has consistently demanded GEDCOM compatibility. Without their support, GEDCOM would not have happened.

In 1989 and again in 1991, at NGS conferences, I held meetings with developers to gather input, inform them of progress, and to review the agreement that held us together. Specifically, I asked if we should continue the de facto approach, or whether they were ready to join us in a more formal process. Both times, the responses were the same as at first, and we continued with the de facto approach.

At the GENTECH conference in 1993, I expressed our continued desire to involve more participants and to follow a more formal process. After this discussion, some mistakenly thought the FHD wanted to give up this activity to others, but our desire is to include more participants in a proven process, through an established standards organization, that has the staying power and recognized discipline to ensure long-term success.

In recent years, Jeri Steele and I independently asked the GEDCOM-L group how we might organize to accomplish this. Only one person responded. His suggestion was “Do it the way the internet does it”.

Needing more explanation than that, I attended the IETF (Internet Engineering Task Force) meetings in December of 1994 and 1995. This is where internet standards are hammered out. I attended the working sessions of the HTML group and others, and discussed GEDCOM and WWW genealogy databases with Tim Berners-Lee, of WWW fame. I met with John Klensin, director of the applications group which created the email, WWW, and other internet application standards, and which would preside over a GEDCOM internet standard.

At the last meeting, John and I agreed that the IETF would not be a proper place to host a GEDCOM standards effort at this time because the people who currently attend IETF meetings do not possess the genealogical expertise to ensure an effective standard. One approach might be to engage a representative set of people to attend IETF and go through their process. Or, John suggested that we contact the Z39 committee again in Washington D.C. Several from his IETF group would be willing to participate in a GEDCOM effort to contribute their standards expertise.

For the present, it has seemed best to finish the long-awaited 5.5 release under the current process, and continue looking for an improved process while waiting to begin on GEDCOM 6.

Bill Harten

GEDCOM Product Manager

January 24, 1996