GENEALOGY

In recent years genealogy has become a spare-time hobby/addiction of mine. However, though I do engage in the self-indulgent pastime of researching my own family, I have found that questions of how to use a computer to assist in genealogical tasks is an extremely interesting source of problems of wider interest to computer science. This happened when a few years ago, in search of guidance as to how to do genealogy 'properly', I investigated what sort of algorithms and heuristics historical demographers were using.

Here is a query I posted at the time - possibly my first message to the genealogy newsgroup which was at the time called soc.roots. Since then I have explored, discussed with numerous new Internet colleagues, and had several computer science students undertake projects related to, the application of computers to genealogy - a subject which can be surprisingly representative of the sort of complex, untidy, dynamic, ambiguous real world that challenges many information system designers.

One such project was on name matching, and resulted in the development of an algorithm that was demonstrably a distinct improvement over Soundex. A paper on this available in PDF form. Another is on a method of Web searching that can cope with the many and varied ways in which people's names might be given in indexes and transcriptions. This is available in postscript form (1 Mbyte) and as an ASCII file (98 kbytes). A third project resulted in the implementation of a means of estimating missing dates in GEDCOM files.

By way of light relief, my personal genealogy interests centre on Carmarthenshire, and North Devon - in particular the village of Clovelly - though I am also researching several lines in Norfolk. The article on Clovelly and the Randells was written with my cousin Francis David Randell and has been reprinted in the journal Maritime South-West; here are two further articles we've prepared that have appeared in the Llanelli Star - Sailing With the Randells, and What Was it Like in 1888 (based on the life of David Randell MP). In researching my Randell ancestors I have collected references to a sizeable number of (often privately) published genealogies and autobiographies, some of which I have already been able to inspect, others of which I am still seeking. I would be grateful for information about any of this latter set (e.g. the date and location of the earliest ancestor recorded).

Turning to the wider world of online genealogical information, in 1994 I was engaged with others in setting up the soc.genealogy.uk+ireland Usenet genealogy newsgroup (since split into soc.genealogy.britain and soc.genealogy.ireland), and am now heavily involved in the development of the associated United Kingdom & Ireland Genealogy (GENUKI) web server, especially the part related to Devon. GENUKI has been a very interesting exercise from a computing science point of view, since what we decided was needed was a scalable distributed web server, i.e. one that would remain coherent and easy-to-use as the number of servers and people involved, and the amount of information provided, grew by several orders of magnitude - as indeed they have, since it now (as of September 2009) has over 90,000 pages. GENUKI is thus now becoming a very useful resource - which I encourage readers to try out. (To my pleased surprise, as a result of GENUKI my colleague Phil Stringer and I have been awarded the Institute of Heraldic and Genealogical Studies' Julian Bickersteth medal, and I have been made a Fellow of the Society of Genealogists.)

In the 1990s I also contributed to Cliff Manis' GenServ Project. (I was editor of the User Manual, but for lack of time have now had to hand that task on.) GenServ is a server containing a large collection of GEDCOM databases (containing over 22 million names, in Nov 2007), submitted by individual genealogists from all over the world. Anyone who has contributed a database to GenServ can perform searches on the entire collection, and receive sophisticated reports on any matching data, all via an automated email service. I would be very pleased to see many more UK-based genealogists contribute their GEDCOM files to this service.

Brian Randell, 26 September 2009