Long long ago in a galaxy far far away I had the theory that I would research the direct lines with little understanding how that would work. Much of that had to do with the available methods at the time. As I became older and (hopefully) wiser and got better at math I came to realize that the so called "direct" lines weren't so direct or distinct. Being a history minor that should have dawned on me sooner, but I'll blame that on having literally zero knowledge of my maternal heritage until I was in my late 30's. The math part comes into question on my paternal side too, since the reverse geometric progression says there wasn't a wide swath of available spouses and the odds of certain folks *not* being related are pretty slim.
Thanks to the internet I've found a fairly large swath of books, both commercially and self published, covering the various branches of my family from various directions. Now I understand the the authors of these aren't infallible, but I don't subscribe to the theory that unless you have a notarized document in front of you it's not true (and old REALLY true if you witnessed the signing). I also understand that often authors use other references as sources, and could propagate an error that was in the original. So be it. The fact is not all of these people are wrong or poor researchers, and when you correlate this with other sources and public records you end up with what I'd call a "preponderance of evidence" proof.
I'll refer to an argument about one of my ancestors. She's referred to as Elizabeth Brooks in two books. One of her sons has been referred to as Michael Brooks Costner. It was common in the time to have the surname name of the mother, or paternal grandmother, as a middle name. Documentation of her father is a tad hazy, and the author of the first book that referred to her Elizabeth Brooks had a few errors in his research. The author of the second book did talk to the author of the first, but also says she referenced the documents in the courthouse in the 30's, but those documents are no longer available. According to purists there is no "proof" her maiden name is Brooks, and have asked that name be stricken from the records on WikiTree due to lack of evidence. Obviously I'm not in agreement with that request.
Since this is not a singular occurrence I've been working to enter ALL of the referred families in several publications prior to 1900. This is not such a herculean effort as it may seem. What I'm finding is there will be a small number of people to enter from the early years, and multiple families are migrating together so there's a large overlap. What this means is source A may have 2500 people, and source B may have 2500 people, but 1500 of them are in source A. Source C also has 2500 people, but 1000 of them are in Source A and 1000 in Source B.
What I've started doing is more explicitly citing my sources as I've been reviewing. For example I may have:
#37 Jane Doe born 1878 married John Smith born 1876 in 1896. John was the son of Simon Smith.
For that I have reference number 37 for Jane Doe, so I enter that as a fact, with the fact source being Doe Genealogy page 100. I also have a name source for Jane Doe, John Smith, and Simon Smith. There are birth sources for Jane and John, and a source for the marriage, and a relationship source for John and Simon. One line with 7 source references.
As there are bits of information in each source that are "filling" such as dates, locations, etc., then other internet searches become more practical since a John Stone is a tens of thousands search, but John Stone in Surry Co, North Carolina in 1810 is a different situation. The RM tools suddenly starting working. That said, much of that is not considered "proof", as there's not a marriage certificate, birth certificate, etc, among them.
I'm trying to enter all the references for each fact to support my "preponderance of evidence" theory. It's a bit of a pain as when I'm working through a book with:
John Doe 100 b 1792
Even though the info may already be in the database, that's three clicks after I've already clicked on the person. Considering that's going to be four more clicks when I try to sync with TreeShare it's starting to get pretty painful. If I was interested in the information only for myself I really wouldn't enter it, but since I'm trying to share this with others I'm unsure if I'm wasting my time or being helpful.