- From: Hamish Harvey <hamish@hamishharvey.com>
- Date: Sun, 19 Nov 2006 13:38:43 +0000
- To: "Arjohn Kampman" <arjohn.kampman@aduna-software.com>
- Cc: "Dan Connolly" <connolly@w3.org>, "Ivan Herman" <ivan@w3.org>, semantic-web@w3.org
On 17/11/06, Arjohn Kampman <arjohn.kampman@aduna-software.com> wrote: > And in case you're looking for name parsing challenges, I'd suggest you > have a look at the following page :-) > > http://en.wikipedia.org/wiki/Jan_Vennegoor_of_Hesselink Being able to parse that automatically might be a challenge. By looking at that page, though, a human could find enough information to decide that his surname is "Vennegoor of Hesselink". They could then encode it that way using the fields of even a parochially English names data model, and expect a display name to be properly reconstructed. It's essentially similar to a "double-barrelled" surname. It seems likely to me that one could never develop a data model and processing software for names which did not fail with some names. One could certainly never be sure one had done so since, as every good Popperian knows, only a single counterexample would ever be needed to prove you wrong. So perhaps part of the key is in having a flexible enough system of manual overrides. For example: given values for a set of fields, we can construct default display and sort forms of a name, but we can also provide these explicitly, in which case the explicitly provided version is used. Even coming close to universality will surely require a load of "cultural modules" in the ontology and the processing code. Cheers, Hamish -- Hamish Harvey Research Associate, School of Civil Engineering and Geosciences, Newcastle University
Received on Sunday, 19 November 2006 13:38:56 UTC