Submitted by testuser42 in this thread: message/view/evidence+vs.+conclusions/30847149 (please read, if interested)
Also available as PDF


I would love to see a (little) more complex example in a pseudo-BG (or DeadEnds or GEDCOM) format, so I could try to diagram that. What will happen when a few more sources and evidence lead to additional information and events? Will new conclusion persons need new conclusion events?


brianjd 2010-12-03T11:15:40-08:00
Evidence and confidence
I notice you didn't rate the confidence of the evidence, but you did the hypotheses. Is there a reason for that? I'd think you'd want to rate the reliability of sources. That would give one way to rank a preferred version of any evidence. I would also think if planning on supported the evidence-hypotheses model we should include a ranking field for any evidence type entity. I'm glad I don't use that model, personally. But, I would defend a person's right and desire to use that model.
testuser42 2010-12-09T11:04:02-08:00
I agree, there should be a possible "confidence" PACT for sources and any- and everything else imaginable. I just wanted to keep the amount of text readable in the graphic.
I also ideally would want to have a way to keep a record of the person who entered/edited data, and a timestamp for the edit. So that if I first trust a certain researcher's opinion I could still change that easily later.
greglamberson 2010-12-09T11:40:47-08:00
Right now we've got to support the variously implemented numeric surety rating system that software applications (and old GEDCOM) have used up to now. I would love for something better and new to develop, and we could certainly insert such a system into our model.

Regarding who entered the data, that is the function of the researcher sort of information that would be attached to the overall database. Timestamps are also something we'll need to add.I think there are several little things like this that we'll need to add onto any model. However, we still don't have basic issues solved, so this hasn't been anything I've focused on.
AdrianB38 2010-12-09T12:41:48-08:00
"Right now we've got to support the variously implemented numeric surety rating system that software applications (and old GEDCOM) have used up to now"

We've got to load up that data somehow, yes. See and its discussion page for my initial stab at a simple advance. If anyone thinks my categories are wrong - please discuss!
GeneJ 2011-01-07T17:29:16-08:00
Discussion of Testuser's evidence-hypothesis model
Hope it helps to set this discussion topic up to hold comments about testuser's illustration
GeneJ 2011-01-07T18:51:15-08:00
Testuser wrote, "used 'hypothesis' since someone said "conclusion" is too final. But 'conclusion' seems to be more common."

As I understand it, what testuser is trying to illustrate is what Tom and others at one time were calling an "evidence-conclusion" approach. While it maybe quite different than the GENTECH model, it doesn't appear so different to me.

My understanding of the evidence-proof-conclusion process is markedly different; my understanding of an evidence based model is markedly different.

Mills Evidence Process Map and the Genealogical Proof Standard do a good job of walking us through how a "proof" is to be developed. --Those are the functions on whichc I focus.

Separately, Board for Certification of Genealogists (BCG), _... Genealogical Standards Manual_ (2000), see p. 25 in "Teaching Standards," as item/standard no. 72, "Database programs accommodate sound data-collection, evidence-evaluation, and compilation standards and do not force or encourage users to leap to premature conclusions about personal identities or relationships, or to tailor their research findings to the input interface."

Without making an attempt to be comprehensive, in the earlier discussions and Testuser's example, I'm hung up on:

(1) Over simplified examples; conclusions don't seem based on sufficient knowledge/information to support a proof.

(2) Apparent emphasis on compilation, and the commingling of "source data" and conclusions.

(3) Lack of consideration for the body of evidence and inferential genealogical practices.

(4) Seems to emphasis collecting research materials rather than the constant evaluation of the information being located. I see this as a "put it off until another day" approach.

(5) Has a "make work" feel to it.