Question question about webtrees possibillities

Hi everyone,

maybe a strange question, but I'll ask it anyway. Lately I've been reading a lot about genealogy and how to set things up "best". Now I'm also reading a lot about the so-called evidence-first type of software, and I have to say I'm very happy with it (a short explanation below).
As far as I know there are only 2 software programs (Centurial and Evidentia). Both packages have their disadvantages (Centurial hasn't been updated for several years now, and both packages are not web-based).

My question is actually whether it's even possible to make something like that in webtrees (with or without a plugin). I'm not a developer by any means, so I'm mainly curious whether it's even possible with a package like webtrees (ultimately everything is possible, but the question is more specifically whether it's possible within reason).

I would prefer to stick with webtrees, but I would really like an input method like Centurial to work, and I also believe that it raises the level of the family tree (at least for me).

Centurial method in a nutshell:
- upload a source, including source reference etc.
- then tag the facts in the image and link them to fields in the database (name, date of birth, parents etc etc)
- link the data to a new/existing person

ultimately it is a different way of filling your database, which certainly ensures that you have the facts at hand very quickly through tagging.

hopefully there is someone here with a bit more knowledge of development and can answer whether something like that is even possible.

I’ve always wondered the same thing about the possibility of creating a different data entry model that creates the Repository_Record first, followed by the Source_Record, then adding facts and citations to Individuals (all of them) last!

To do this (I’m not a PHP programmer) two data entry screens need to be added the Repository and the Source.  This would be easy for someone to create.

The hard part is the, as you call it, “source reference”, which I assume is the citation information like page number for a book, page/city/district for a census, url and date observed for a web page, because these are only added to a GEDCOM when you actually add a fact to an individual and that individual must already be in the GEDCOM/database.  The database needs to be able to add facts to a “stated person” but not necessarily a person that should be in your database.

First, GEDCOM is not very well normalized, so any database design must have facts and citation information not imbedded in the GEDCOM Individual Record.

Second, Many people and Software programs do not understand the difference between a “Source” and a “Citation” so everyone enters data differently into their databases, some call a source the book, census or website, other define a source that includes the page, census page, website URL.  

Third, most software does not have a good way of identifying and maintaining fact data that is in conflict with other fact data.  For instance: two different sources assert different birth dates.  During collect of data, these two assertions need to be entered with equal possibility of correctness based on the collected evidence. While it is true that you can eventually set a reliability value on the assertion, this is not something you can do when entering the evidence in this model.

Fourth, the stated person in an assertion may not be someone in your family, for example you find a John Doe in a census that could a grandfather with parents of Bill and Susan, you don’t know if this John is your grandfather and you at this time don’t know the names of your great grandparents.  Entering the Repository and Source is fine, but then you have to add Bill and Susan Doe (to record their data) but you are only adding potential family info.  Normally, I would collect the data in a separate file (not webtrees) then look for other assertions from other sources, until I have multiple assertions that raise the potential of a good link.

 

I did think about this a few years ago, and created a basic design and simple prototype.

My approach was to create a mini tree/GEDCOM file for each source.

So, a baptism record might create a tree containing 3 individuals, a birth for the child, a marriage for the parents, occupation for the father, maybe birth/death events for the parents with BEF/AFT dates, etc.

This are my "evidence individuals" - all in GEDCOM format.

I'd then create a "conclusion individual", which contained links (GEDCOM ALIA tags) to the "evidence individuals", and also notes to justify my conclusions.

The advantage is that if you make a mistake in your conclusions (e.g. you have assumed that two records are the same individual when they are not), it is easy to "undo" the conclusion and create a different conclusion.

Greg,

If I understand you correctly, you would create an "Evidence Individual" for each source. 

1) The Marriage would create 2 partner individual records. i.e. Snow White and "The Prince"
2) A source indicated that "Mrs Prince" had a child, fathered by "The Prince"  3 additional Individuals would be created, (we can't be sure it is the same "The Prince"! (child, Mrs. Prince, The Prince)
3) A Baptism happened (3 additional Individuals are created)
4) Other events occur to people that generate additional Individual Records, could be the same three or unrelated people.

When we Conclude they are the same people in any instances from above a "Conclusion Individual" is created for each of the three individuals that have no facts, but points at (via ALIA) to the appropriate "Evidence Individuals".

Do you at any point bring over the actual facts from the "Evidence Individuals" to the "Conclusion Individual"?

 

 

fisharebest wrote:

I did think about this a few years ago, and created a basic design and simple prototype.

My approach was to create a mini tree/GEDCOM file for each source.

So, a baptism record might create a tree containing 3 individuals, a birth for the child, a marriage for the parents, occupation for the father, maybe birth/death events for the parents with BEF/AFT dates, etc.

This are my "evidence individuals" - all in GEDCOM format.

I'd then create a "conclusion individual", which contained links (GEDCOM ALIA tags) to the "evidence individuals", and also notes to justify my conclusions.

The advantage is that if you make a mistake in your conclusions (e.g. you have assumed that two records are the same individual when they are not), it is easy to "undo" the conclusion and create a different conclusion.

Sounds great, i even think i found something on github about this project, is that possible? (ofcourse i can't find it now).

I believe this is in basic "the right way" when doing data entry in a evidence based way. 

Forgive me my rude question, but what did you decide to stop developing this project? didn't you believe in the project? or where there other reasons? (and how can i ignite the flame again )

norwegian_sardines wrote:

Second, Many people and Software programs do not understand the difference between a “Source” and a “Citation” so everyone enters data differently into their databases, some call a source the book, census or website, other define a source that includes the page, census page, website URL.  

I believe that with an evidence-first software application, the quality of the input of sources/citations also increases.
What you see with the "conclusion-first" software is that adding a source in many cases always contains the same package of information.

A gravestone requires different source metadata than, for example, a census. In fact, a photocopy of a census on a website has a different source reference than a census that you have viewed in hard copy in an archive somewhere.

By placing the emphasis on evidence first, the basic data of the sources will improve considerably. In an ideal world, the software would
 

In fact, a photocopy of a census on a website has a different source reference than a census that you have viewed in hard copy in an archive somewhere.

I don’t understand your term “source reference”, please explain! I assumed it was the “where found” within the actual source, but this statement does not jive with that use!

what i mean with another source reference is that when i find something online i use the online resource, after all that is where i find my documentation. then i need information like a url, when consulted etc.

when i find the same source but then the original in hardcopy, then i want to have a storage location, book number, page, etc etc.

in the first example not all desired information is regularly available.

little note: i'm not a source expert, and as you said in your first reply in this topic, most people don't understand sources/citations well. (I'm for sure one of them). so having software that is helping me "doing the right thing" will be great for sure!

Wouldn't a setting that makes it mandatory to add a source citation for each fact you add be a reasonable compromise?
That way you can at least enforce sources for every fact, and I guess it wouldn't be so complicated to support in webtrees.
 

In my humble opinion it is more about the philosophy than just forcing to add a source.

I think that an evidence first approach will make you look at sources in a different way.
More like deep dive into the source and retract all the valuable information instead of adding a source and only take out the information which will prove your statement.