What companies I talked with at RootsTech said about the mission of FHISO.
First let me say that I do recognize that this post is a bit overdue. Overdue because I probably should have posted at least once in the past 70 days, and overdue because the events on which I am reporting are now several weeks old. However, I find that almost all of my various travels occur in February and early March, meaning I have to spend January getting my classes ready and all of my spare time keeping up with the needs of the few percent of my over 500 students whose lives are making the normal course of my courses not work well for them.
But now it is spring break, and I have time to write. And what shall I write? About what people at RootsTech seemed to want in information standards.
The Family History Information Standards Organisation (FHISO Commonly pronounced /ˈfaɪzəʊ/ ) is broadly interested in creating standards for (mostly computerized) information formats related to family history. Being a representative of FHISO, I spent much of RootsTech wandering around to various company’s booths on the expo floor hall asking them what standards they wish exist and which of the ideas FHISO has discussed they care about.
GEDCOM is arguably the only standard ever to reach the family history market, and has excited distaste since its inception in everyone from its designers down to the least technical people around. But it was still the most discussed, most-requested topic I broached and the most common one to have others bring to me unsolicited too.
Almost every company that I talked to used GEDCOM Most couldn’t name the version number they used; those that did used either 5.5 or 5.5.1. , although almost all used it as an external format only, it not sufficing for what their own tools needed. Many people requested nothing more than an update to that standard, an agreement on what the set of additional concepts should be added to the standard. Most companies have already added their own extension tags, but recognize that it would be nice to all use the same set of extensions.
I noticed without surprise or judgement that companies that focussed on GEDCOM in our conversations tended (with some exceptions) to fit into one of two categories: either their main data model was human- instead of machine-centric with lots of free-form text; or the company had a somewhat static feel, happy with their niche in the market and not discussing new directions. However, the newest companies (those started in the past two years or so) generally had little to no interest in GEDCOM, some of them not having even tried to implement it. Whether that indicates lack of maturity or a sign of things to come I do not know.
FamilySearch created GEDCOM, and recently rolled out a new specification called GEDCOM-X. Because FamilySearch requires all of its partners to communicate through GEDCOM-X, and because most companies find it worth their while to work with FamilySearch, I asked each company I talked to about GEDCOM-X.
To my surprise, the most common reply was that they had only a partial implementation, only enough to parse the particular material FamilySearch was sending them. This was often followed by a chance to gripe about whatever was annoying them about GEDCOM-X at the moment. These were not offered as serious setbacks, but they did convey a mood. Most complaints were either about GEDCOM-X’s verbosity, about the verbosity of the HATEOS design principle used by the FamilySearch API, or undefended but deep-set fears that GEDCOM-X and/or the API would not be stable enough to be worth implementing. A minority of complaints centered on the GEDCOM-X specification’s format which some people found difficult to follow.
When I raised other topics, everyone (I think without exception) was happy with a suggestion that we would standardize digital citations. However, the moment I went past that big idea, consensus vanished. We all know that that citations are hard to share now and feel that they thought this should be fixable… somehow.
I was able to find someone on each side of the following options:
Should “Citations” be unicode text, or text with italics and other formatting, or sets of unordered key-value pairs?
Should the keys (e.g., that we store author, title, etc.) be open-ended, or based on a set of defined templates, or based on a single large/small set of terms, or some kind of mix?
Must a citation contain sufficient information to locate the described source physically?
Must each pair of citations contain sufficient information to determine if they are talking about the same source?
Should citations be able to refer to other citations?
My overall impression is that, as a community and with many individual exceptions, family history has not thought carefully about citations and what we want from them.
I was unable to find any other topic that interested more than a few people. Often when I would say to one person, “I was just talking to someone else about X” I would get the emotionless stare of someone trying to decide if being nice was sufficient motivation to feign interest in that topic. This is not to say that people disapproved of one another’s goals or would fight people trying to add those extra standards, just that they were not particularly motivated by them.
More RootsTech and Family Search Develop Conference reports will follow in my next few posts.