gramps-devel

 [Gramps-devel] probably_alive's ranges cause confusion From: Douglas S. Blank - 2007-01-16 06:23:44 ```So, I'm trying to understand what doesn't work correctly with the filter ProbablyAlive, when I found some issues with the current values in the function probably_alive. I wanted to bounce them off of you here, before filing a bug and fix. Currently, there are some range values in probably_alive that make sense when thinking about a single person, but applied recursively, it makes a confusing situation for the user. The idea in not_too_old and too_old is that they look for things within 110 years, and beyond 150 years from a given year. However, when these are used on the generation before the current one, you can get the following results. Suppose that you have this data: person, lastname, firstname, birthdate 1, Smith, John, 1600 2, Smith, Joe, 3, Smith, Fred, marriage, father A, 1 family, child A, 2 A, 3 (You can import this from the Spreadsheet plugin, if you want to follow along at home.) If you apply the "Probably alive" filter on these years, you'll get these results: 1600-1729: All alive (conservative, but ok) 1731-1749: Father alive, sons dead (confusing) 1751- : All dead The problem lies in those fudge factors in not_too_old and too_old as applied to the current person. If we change the 110 and 150 to the same value, then we have more consistent results. Here, I've set all of the 110 and 150 values to 100 to make it easy to see: 1600-1699: All alive 1701-1719: Sons alive (correct, not confusing) 1721- : All dead One hundred years is a lifespan, and twenty years is the average generation years (which is why the sons are assumed dead in by 1720, eg 1600 + 100 + 20). This is also a problem here. We should consider that the parents can begin producing offspring after they turn 20, but they can continue to do so for, say, 30 more. So, we should consider the sons possibly alive until 1750. The 150 value was, I suspect, there so that if there were events (maybe a will-signing) that happened a few years after the death, we could account for that. But this causes the above inconsistent issue. The 20-year is a bug, because it should consider the end of the productive years, not the beginning. If you all agree, I'll submit a patch to fix these two issues. And then head onto the ProbablyAlive filter. -Doug ```