Related Work That Got Lost In the Cracks
I’ve made a lot of notes over the last ten months. Stacks of ideas, sketches, lists, references, and more. I was just going through some of them when I noticed a handful of online sources that I had considered citing in my thesis proposal, but completely forgot.Â
- GenWeb and Family Web are both loose specifications for recording genealogical information in a semi-structured manner with only rudimentary HTML formatting elements. GenWeb was first proposed in April of 1994 and subsequently worked on until August of 1997. Family Web was independently invented by David Croft in late 1996 with increasingly sporadic activity through April of last year. In both schemes, each Web page corresponds to a single individual, with hyperlinks to denote family relationships.  These links provide a natural mechanism for distributing the genealogical data across multiple servers. These specifications have the compelling advantage of being very simple to understand and author.
- The Semantic Web for Family History is a Web page that discusses how RDF can be used for genealogical data, along with advantages and drawbacks to this approach. I stumbled across this when I was first toying with the idea of using RDF for genealogy. It served to reinforce my ideas and suggest problematic areas that I should focus on, such as how to handle false or potentially destructive data.
If you’re aware of any other similar efforts, please let me know. It’s far too easy to miss something in such a big world.
Technorati Tags:
Filed in Thesis
Comment feed for this post
Many of your ideas for the Distributed Family Tree have already been implemented in the PhpGedView Open Source genealogy project. In PhpGedView you can link up people across remote websites using web services. Data is automatically and seemlessly transferred between sites effectively joining the trees without duplicating the data.
The idea for this was first introduced at the 2005 FHT conference. You can read the paper here:
http://www.fht.byu.edu/prev_workshops/workshop05/FHTCD/session4/s4-JohnFinlay_GedcomCGIProtocol.pdf
and view the slide presentation here:
http://www.fht.byu.edu/prev_workshops/workshop05/FHTCD/session4/s4-FinlayGedcomCGIGDBI.ppt
Since presenting this at the FHT conference we have implemented the idea in PhpGedView v4.0. You can read about how we implemented it here: http://wiki.phpgedview.net/en/index.php/How_Remote_Linking_Works
-John
Wow John, this is really excellent! Thank you for bringing it to my attention.
It was previously my understanding that PhpGedView was only a way to publish GEDCOM files on the Internet so that they could be interactively browsed and edited. I was unaware that installations of PhpGedView can remotely link to each other.
I would really like to use this protocol for my thesis. I will need to make a few extensions to accomodate the data model that I’m proposing, but I think the two can interoperate quite nicely. It would be quite silly to create a distinct genealogy network when a satisfactory one already exists.
The remote linking feature was just added with the release of PhpGedView version 4.0 earlier this summer. So it hasn’t been in production for a long time, but it has been under development since spring 2005. It took a while to get it out because we had to also develop the web services, etc. to handle the connections.
You are welcome to use as much of our work as you want in your thesis. The protocol should be fairly extensible to meet your needs. Just let us know about any changes you make so that we can keep compatibility. The Genealogy Web, or Distributed Family Tree, will only work if things remain compatible.
I am looking forward to seeing the web service APIs released by the LDS church that there has been so much talk about to see if it would be possible to remote link to them. Still waiting for the APIs to be released though.
The next phase that I would like to start is a global search. Right now you can search across multiple sites from another search page in PhpGedView, but you have to manually enter in all of the sites. Unfortunately PHP is not multi-threaded so it takes quite a bit of time for all of the web-services to search all of the remote sites. There are all sorts of considerations for actually implementing the global search such as caching and playing nice with everyone elses websites.
I see that there is a directory of PhpGedView sites at http://phpgedview.sourceforge.net/registry.php. Is there any way to access an XML version of this directory? I’m thinking that a good place to start development on Genesis is to write a simple client that can navigate the Genealogy Web. A machine-understandable directory of PhpGedView sites would really facilitate that.