- Posts: 25
Question Wrong birth calculation
- herver
- Topic Author
- Offline
- New Member
Less
More
2 years 10 months ago #1
by herver
Wrong birth calculation was created by herver
I found a possible bug in the estimated birth estimate. I have the case of someone with no date of birth that webtrees calculates was born in 2028. His father has loaded as date of birth 1984, his mother does not have date of birth loaded but he has to get married in 2013. If you tell me that other information I should send to find the error I do it with pleasure. I am using webtress 2.0.16
Thank you
Greetings
In Spanish
Encontre un posible bug en el calculo estimado de nacimiento. Tengo el caso de alguien sin fecha de nacimiento que webtrees calcula que nacio en 2028. Su padre tiene cargado como fecha de nacimiento 1984, su madre no tiene cargada fecha de nacimiento pero si tiene que se caso en 2013. Si me indican que otra informacion deberia mandar para encontrar el error lo hago con gusto. Estoy usando webtress 2.0.16
Muchas gracias
Saludos
Thank you
Greetings
In Spanish
Encontre un posible bug en el calculo estimado de nacimiento. Tengo el caso de alguien sin fecha de nacimiento que webtrees calcula que nacio en 2028. Su padre tiene cargado como fecha de nacimiento 1984, su madre no tiene cargada fecha de nacimiento pero si tiene que se caso en 2013. Si me indican que otra informacion deberia mandar para encontrar el error lo hago con gusto. Estoy usando webtress 2.0.16
Muchas gracias
Saludos
Please Log in or Create an account to join the conversation.
- ungeahnt
- Offline
- Junior Member
2 years 10 months ago - 2 years 10 months ago #2
by ungeahnt
Dieter
Schmidt ⚭ Schwab (Sudeten) | Hauer ⚭ Bühler (Bayern / BW) | Маринов ⚭ Шаламанова (България)
webtrees 2.1.17 | PHP 8.1 | MariaDB 10.5
Replied by ungeahnt on topic Wrong birth calculation
Hi herver,
as far as I know, the calculation of estimated birth/death dates was implemented mainly for the internal sort order in lists. The calculation should be performed as quickly as possible and without requiring large amounts of server resources. One simply wants to determine an approximate date. Accuracy was not so important to the developers.
The display of these values in wt is only an additional feature (like a free addition). I tested the estimated data some time ago (wt1.7), but then decided to turn it off again (= default setting). The automatically determined values were simply too much off. More serious, however, is the fact that you can not see at first glance if the data was estimated by the machine or a human! A human achieves significantly better estimates than a quick and dirty algorithm. My recommendation is therefore not to use this function for display.
Then to the actual question:
I can reproduce the date of birth in the future and that makes no sense. An additional check to prevent "future events" would certainly be useful. However, each additional check requires more resources on the server and this is not desired for the calculation of estimated data in general. But maybe the additional overhead is not that big in this case?
You can create an issue on github and maybe a solution can be found.
as far as I know, the calculation of estimated birth/death dates was implemented mainly for the internal sort order in lists. The calculation should be performed as quickly as possible and without requiring large amounts of server resources. One simply wants to determine an approximate date. Accuracy was not so important to the developers.
The display of these values in wt is only an additional feature (like a free addition). I tested the estimated data some time ago (wt1.7), but then decided to turn it off again (= default setting). The automatically determined values were simply too much off. More serious, however, is the fact that you can not see at first glance if the data was estimated by the machine or a human! A human achieves significantly better estimates than a quick and dirty algorithm. My recommendation is therefore not to use this function for display.
Then to the actual question:
I can reproduce the date of birth in the future and that makes no sense. An additional check to prevent "future events" would certainly be useful. However, each additional check requires more resources on the server and this is not desired for the calculation of estimated data in general. But maybe the additional overhead is not that big in this case?
You can create an issue on github and maybe a solution can be found.
Dieter
Schmidt ⚭ Schwab (Sudeten) | Hauer ⚭ Bühler (Bayern / BW) | Маринов ⚭ Шаламанова (България)
webtrees 2.1.17 | PHP 8.1 | MariaDB 10.5
Last edit: 2 years 10 months ago by ungeahnt.
Please Log in or Create an account to join the conversation.
- herver
- Topic Author
- Offline
- New Member
Less
More
- Posts: 25
2 years 10 months ago #3
by herver
Replied by herver on topic Wrong birth calculation
Thank you very much for your answer. I have the estimate active because it helps me to have a quick snapshot of when a person died or was born and to be able to have a quick estimate when I am looking for ancestors. But it was the first time that I came across an estimate towards the future.
I'm going to investigate the subject a little more to make the report on Github, since I don't know how the calculation works. Perhaps you can implement a system that does not calculate all the time, but is a specific task to perform every certain amount of time and check by human, in the style of Data fixes
Greetings and thank you very much again
I'm going to investigate the subject a little more to make the report on Github, since I don't know how the calculation works. Perhaps you can implement a system that does not calculate all the time, but is a specific task to perform every certain amount of time and check by human, in the style of Data fixes
Greetings and thank you very much again
Please Log in or Create an account to join the conversation.
- ungeahnt
- Offline
- Junior Member
2 years 10 months ago #4
by ungeahnt
I think its starting here:
github.com/fisharebest/webtrees/blob/6a9.../Individual.php#L585
It should be one of the getmarriageDate() -> 30 years / 2 = 15 years: 2013 + 15 = 2028
Dieter
Schmidt ⚭ Schwab (Sudeten) | Hauer ⚭ Bühler (Bayern / BW) | Маринов ⚭ Шаламанова (България)
webtrees 2.1.17 | PHP 8.1 | MariaDB 10.5
Replied by ungeahnt on topic Wrong birth calculation
I'm going to investigate the subject a little more to make the report on Github, since I don't know how the calculation works.
I think its starting here:
github.com/fisharebest/webtrees/blob/6a9.../Individual.php#L585
It should be one of the getmarriageDate() -> 30 years / 2 = 15 years: 2013 + 15 = 2028
Dieter
Schmidt ⚭ Schwab (Sudeten) | Hauer ⚭ Bühler (Bayern / BW) | Маринов ⚭ Шаламанова (България)
webtrees 2.1.17 | PHP 8.1 | MariaDB 10.5
Please Log in or Create an account to join the conversation.
- ungeahnt
- Offline
- Junior Member
2 years 10 months ago #5
by ungeahnt
Dieter
Schmidt ⚭ Schwab (Sudeten) | Hauer ⚭ Bühler (Bayern / BW) | Маринов ⚭ Шаламанова (България)
webtrees 2.1.17 | PHP 8.1 | MariaDB 10.5
Replied by ungeahnt on topic Wrong birth calculation
btw:
" free addition" sounds funny for an open source software... is only an additional feature (like a free addition).
Dieter
Schmidt ⚭ Schwab (Sudeten) | Hauer ⚭ Bühler (Bayern / BW) | Маринов ⚭ Шаламанова (България)
webtrees 2.1.17 | PHP 8.1 | MariaDB 10.5
Please Log in or Create an account to join the conversation.
- herver
- Topic Author
- Offline
- New Member
Less
More
- Posts: 25
2 years 10 months ago #6
by herver
Replied by herver on topic Wrong birth calculation
Please Log in or Create an account to join the conversation.