From: Duncan L. <du...@li...> - 2006-04-21 13:35:46
|
On Fri, 2006-04-21 at 07:23 -0400, Jim Winfrey wrote: > Hi Duncan, > > Early on I used the estimated dates feature in FTM. I have spent > years fixing the errors it caused. There is just so much error in old > records that we have no idea when people were born, whether there may > have been a later wife/mother we don't have a record for, even whether > the child shown on the census belonged to the subject family. I would > want to be able to control how, and when, this feature was used. I > find it just as easy to make this determination manually. > > Jim Certainly it should be tagged in a specific way, so it's clearly auto-generated. But if it was smart and optional I could imagine using it. Imagine if you could teach it average lifespans in a certain decade, it checked all events to make sure they're alive and old enough for them. I've marked unknown births as "estimated before nnnn" with a comment that it assumes they were at least 15 when they had their first child. Likewise marriage dates can help suggest other dates. Is it worth making as a suggestion for gramps? Duncan |