Could the gedcom import please add references using the advanced notation

+8 votes
160 views

 

So references added gedcom seem to be of this form

'''Samuel''' Hurved. <ref>Source: [[#S48]] </ref>

<references />

* Source: <span id='S48'>S48 Citation</span> .....

but the preferred way to do it is like this

'''Samuel''' Hurved. <ref name="S48">S48 Citation</ref>

So why are they done this way in Gedcom, please its very frustrating.

Ive tried editing people Ive added via gedcom to use advanced sources, and its a pain. I usually find its easier to remove the biography and redo it, it shoudn't be this way. Ive also tried adding people directly into wikitree, this is even worse because then you manually have to add all the relationships and then edit the bio

So Ill be sticking with gedcom for the majority of the tree, and Ill edit the gedcom before importing to simplify those long winded Ancestry sources with their copyright, and Ill remove references to files on my hard drive. But I wont be able to get it to use advanced sources, but I dont see a programmatic reason why this couldnt be done in the gedcom import code.

 

 

in WikiTree Tech by Paul Taylor G2G6 Mach 1 (13.7k points)
Great idea. Upvoted!
Paul,

It's really unfortunate that the recommended method for manually creating repeated references to the same source doesn't match the method using in GEDCOM import. But we couldn't find a way to make them consistent.

Named references like "<ref name="X">" enable you to repeat the exact same reference.

In GEDCOM import, linked referenced do something different. They link to a source but put different content in the reference itself.

There's a help page or G2G discussion that goes into this in more depth but I don't have it handy. Sorry. Again, I know this is confusing and it is unfortunate.

Chris
HI Chris, but certainly in my gedcom they link to a source but put the same content in the reference itself. For example a Census record maybe used to cite name, birth, place of birth in each citation they are referring to the same source and same reference each time.

I spent yesterday/today seeing what I could and I have made some progress. Ive created an opensrc project  https://bitbucket.org/ijabz/wikitreebiographer that can be used to create biographies for people in a gedcom, an example of the bio they create is http://www.wikitree.com/index.php?title=Hurved-108. Currently its quite basic and doesn't include events such as marriage but the key thing is it does create advanced citations. I need to do some paid work now but will come back to this next week, what do you think ?

Looks great. Thanks for doing that open source! You should think about joining our WT_Apps Project, Paul.

Thanks, yes I'll do that
Chris, it would help me if you could find the discussion page that you mention above.
Ive added introduction to the googlegroup

1 Answer

+1 vote
I have just been down this path, after tidying up Gedcoms and sources first, and expending esubstantial ffort getting relatively few profiles sensible. (300 down, 1400 to go)

But do not underestimate the difficulty of processing a vast range of Gedcoms. Even if they look similar in a variety of applications, under the hood, they have all used different combinations of tags.

Maybe we should start a free-page resource or similar to capture all the quirks and recommended best practices. Maybe there is one!
by Chris Little G2G6 Mach 2 (30k points)
Good idea, I have a thought about an interim solution that could improve things. The Gedcom import does two thing

1. Create people and the links between them

2. Creates a bio with details and sources for each person

It does stage 1 okay, stage 2 is the problem.

I am looking at writing a program that takes a gedcom tree as import and then spits out a wikitree biography for each person. If the biographies were good then after importing the gedcom tree you could simp[ly copy and paste each bio for each person. This would be a bit laborious but less laborious then actually editing each biography.

I will start with the gedcom created from FamilyTreeMaker synced form ancestry because this is what I use, but I could opensrc it so it could be amended for different gedcoms.
That souns a very good idea. Though I suspect that is what the existing import mechanism does.

 

In the Little project we discussed possible Biography structures, and we liked:

Biography, textual, with refs

Timeline, with refs. For birth, death, marriage, child, residence, other events. A very useful view to spot discrepancies, missing sources? I guess this is what your software would attempt to spit out.

Notes - these seem to come through Gedcom mangling quite well.

Sources, but need to think about associated imagery of, say, original Census entry or Will, and any Transcript.(I am trying to read a 1608 manuscript will!)
Ive started work on a project https://bitbucket.org/ijabz/wikitreebiographer - the main aim so far has been on using advanced citations - example output at http://www.wikitree.com/index.php?title=Hurved-108

Any advise, comments or help would be much appreciated.
Paul, that looks a great start of automatically giving a framework for a story and sources.

The next challenge would as you said: marriage, and pulling in details of children into the timeline.

All the references to Java, git, GitHub, pull, push, fork, clone, merge and maven making me feel very out of date!

Some typographical nit-picks, though wiki and markup may limit things that you can do:

1. How about colons after each event title. Like this:

Born:

2. Consistent single space after commas, like, this.

3. Can the consecutive commas be removed, , ? I've spent a lot of time doing this manually. Definitely worth a few beers!

4. Could each of the event lines or source lines end with a full stop? A la Algol! A proper scope delimiter.

HTH, Chris
Thanks Chris I've made your suggested chnages if you would like to check, although Im not sure about the value of the full stops for such short sentences.

That looks really good. I understand concern over full stop on short string, but I was a Grammar School boy! And a programming career made me pedantic ;-)

Is it easy to run first string on from date, to save screen space?

As in:

02 APR 1837: Sandford, Devon, England. [8]

Though I suppose this would increase the chances of a long address flowing onto the next line.

Chris

Well there is no guarantee that you would have date and address. I think for now I would rather keep it simple and concentrate on extracting info rather than worrying about formatting too much.
Good idea. Thank you for the work.
Got something working on marriage, have you an example of a profile showing children Im not quite sure why children are necessary as they are linked in the tree anyway and unlike marriage there isnt a particular event to be recorded involving the parent and child.
The birth,  and death, of children fill out the lifeline and give other useful information like residence or a pattern. The pattern also gives confidence, or the opposite, in the evidence. Lots of children appear every two or so years. Helps identify false assumptions. Recently saw a putative tree where mother was giving birth at 75. Sometimes two children baptised at same time, but not twins - signs that the local church/registry not functioning, or was a considerable distance away, or the vicar charged to much, etc. Sometimes, two children with same name, a few years apart - sign that the earlier one died in infancy, and need to research the death.   

Presenting the same information in different ways Is often very useful. The traditional paper based tools for genealogy included: Pedigree Chart, Chart of Descent, fact sheet (individual timeline) as well as the tree itself.

Chris
Oops, forgot - here is an example. Try Little-4639. Sorry about the preaching.
Chris, an example showing references from an imported gedcom tree would be better, this is very much a handwritten example and hence not so easy to consider. But in my view if I have a record showing a baptism of a child with the person listed as father than this could go in, but a birth record with no reference to the father probably shoudnt go in because then I am just duplicating records that are already on the childs bio.
I am afraid I had manually edited out most of my auto created Gedcom imports, to something that adhered to the advanced source style.

I chose that profile as one that shows both timeline and narrative.

I think your analysis for not including un-parented child is right. I imagine the use case where the profile manager expects it to appear, and is then obliged to correct, if appropriate, the child' profile.

Related questions

+6 votes
2 answers
0 votes
1 answer
83 views asked Jul 12, 2019 in WikiTree Tech by David Dawe G2G Crew (540 points)
+2 votes
1 answer

WikiTree  ~  About  ~  Help Help  ~  Search Person Search  ~  Surname:

disclaimer - terms - copyright

...