Talk:Rodovid/Specifications

Specifications

• Goals - A genealogy wiki for the Wikimedia Foundation will have requirements both lesser and greater than those of a generic genealogy site. There will be a need to limit the scope to a level that is manageable for the Foundation. That means that personal photo albums, for instance, will not be stored on a Wikimedia genealogy wiki.

• To the greatest extent possible, ethnic biases are to be excluded from the data model. A distinction should be made between biological parents and nurturing parents, while recording both kinds of relationships with full respect. No automatic assumptions are to be made about parents or spouses (i.e. as to numbers or sexuality).

• The best model for this type of wiki is probably Wikispecies - a tightly constrained taxonomy wiki. A Wikimedia genealogy project will exist to reference information that is dealt with more extensively elsewhere. In the case of notable individuals it may be that the reference will go to Wikipedia.

• Individuals in this wiki will be defined by means of all recorded data: names, labeled dates, geographic locations, relationships, and verifiable sources.

• Individuals will be tracked internally by a unique identifier, such as a hashed set of letters/numbers, but that will always be represented as a link when presented to the viewer. Ideally, the internal identifier would not be generally available to the public.

• There will be a standard set of relationships but there should also be a moderated capability of adding any verifiable relationship.

• When people are added, there will be an automatic check for other people with similar data, with the goal of minimizing the duplication of people.

• Relationship data will have both a start date and an end date, where known.

• Events will be like relationships except that only one date/time will be recorded.

• It should be possible to reuse the same sources over and over for various individuals - a genealogy sources "common".

• Data uploaded from a file, such as GEDCOM data, will be filtered for relevance to the Wikimedia genealogy database and reviewed by the uploader in a sandbox prior to application to the public database. The data in the sandbox will be flagged as to whether it is new people, new sources, likely to be updates to existing people, and any other relevant issues that may need to be addressed.

• No living person is to be recorded in the genealogy wiki. No deceased person is to be recorded who has a living child, spouse, or ancestor (biological or otherwise). (what else?) Update checks are to be put in place to flag these conditions.

Bob 02:38, 12 December 2006 (UTC)

I'd like to disagree with your last criteria. I think you can put living person, but with restricted data. For example, date and place of birth should not be included, also exact address and city should not be included. And deceased person should be included without exception, and with full data included (date of birth, date of death, etc) 121.6.116.120 06:42, 5 May 2007 (UTC)

---

Here are my comments on the technical specifications:

• I think the basic idea should be that the genealogy data is presented in the context of a Wiki page so that the genealogy information appears like an Infobox control. Inside the new "Genobox" control would be several defined areas. The top area would be divided into left and right halves, with the left half listing people who stand in a male parental relationship to the individual and the right half listing people who stand in a femail parental relationship to the individual. These lists would be hyperlinks for the names of these individuals, with perhaps a parenthetical phrase for the specific parental relationship ("Natural" would imply biological or genetic parent; "Adoptive" would imply an adoption; "Step" would imply a marriage to a biological parent; etc.).
• In the proposed "Genobox" control, underneath the parental area would be the basic identifying information for the person who the particular Wiki page is about in the left side area. Name, date of birth, and date of death would be the minimum information for this part of the box, but other basic information could be included if it was convenient and did not make the area of the box too big. On the right side would be a list of people with whom this person had some sort of a spousal or domestic partner relationship with some brief indication of the nature of the relationship in parentheses (like "Wife" or "Husband" in the usual case). Places of birth and/or death and a variable list of facts about the person associated with dates during (or after) their lives would be in a full-width area underneath the above and above the child and sibling information (below). A single photo of the person could be allowed at this point.
• After the facts about, and photo of, the person, lists of siblings and children would appear, each presented in a manner similar to that described for the parents (above). In fact, perhaps the easiest way to do this would be to have a left box of children and a right box of siblings, following the pattern of male and female parental figures above the person.
• To the extent that any fact is "sourced" in the GEDCOM file (or otherwise), the source would be linked similar to a reference; and in fact, it could be a reference in Wikimedia code.
• If the above is adopted, it should be fairly easy to write code to import a GEDCOM file with listed sources and convert that information into a set of "Genobox" controls placed onto otherwise blank pages.
• Anything within the GEDCOM (or other) file which did not easily fit within the above structure could be converted to text and placed in the main area of the page under a heading like "==GEDCOM DATA==" (or some such). The main text area could be edited with any biographical information appropriate to the person, similar to a Wikipedia article.
• The usual rule is that no person who was born less than 100 years ago is displayed unless that person is dead. The data can still be entered into the database and viewed and edited by the author or an editor with privileges, but it is not displayed to the general public. That person's name would still appear as a relative on the appropriate pages, but clicking on their hyperlink would get to a data notifying the reader that the data is protected. There would obviously have to be some control on who is (or is not) allowed to edit the data for anybody who does not meet the display criteria. That would be my proposal, anyway.
• Automatic merging of two people should be based only upon a high level of certainty that the people are the same. Obviously, manual merging should also be possible. A reverse merge (a "duplicate and split" function, actually) should also be possible, but it will be up to the person performing that task to delete the inappropriate content for each of the two split records. For the merging and splitting to happen, each person in the database must have an arbitrary identification such that even identical records would be kept separate (which would be exactly what would happen as soon as a "duplicate and split" function was executed and before any editing on the results had been performed).

Anyway, that is off the top of my head. I'm sure I will think of some more before I get too far down the line here. ElbonianFL 19:24, 22 November 2007 (UTC)

Tools

• Family Tree
• Full family tree
• Tool to identify relationship between two individuals
• Tool to import genealogy files in various formats, including (at a minimum) a recent version of GEDCOM.
• Tool to manually (or even automatically) merge two distinct people by combining all of their associated data.
• Tool to manually un-merge one person's data by duplicating it into a separate record that could be edited.