I could answer the question in the title of this post very simply; I would add any and all information I have about every single ancestor. Family history is not about any particular quantity of data, it is about adequately and as completely as possible documenting the lives of our ancestors. I guess the question arises in my mind as a result of an experience I had recently that would seem to be totally unrelated.
After returning from #RootsTech 2014, I received an email from a blog reader who was complaining about a problem she was having with MyHeritage.com. She was receiving an excessive number of SmartMatches from an entry on her family tree. Without going into any details, the root of the problem was the way the information was being entered into her family tree. Apparently, she had entered the information as "Mrs. John Jones" using the title "Mrs." as part of the name. The issue was quickly resolved through an extraordinary effort by MyHeritage.com. But the problem raised by the inclusion of extraneous information in a name field applies to all online family trees including FamilySearch.org's Family Tree.
One example that has been around since the early days of Personal Ancestral File (PAF) is the issue of putting LDS blessing information into a data entry for "christening." it seems obvious, that there are many important differences between an LDS blessing and what is meant by a "christening." However, because the field exists, users commonly used that field to enter LDS blessing information without any further qualification. Usually, this was not a problem but in some instances there would be a question as to whether or not the individual ancestor was christened in another church before joining The Church of Jesus Christ of Latter-day Saints. In short, there is a real question about not only what information should be included in Family Tree but also how that information should be entered.
The real issue here is data integrity. I would maintain that entering a name using the format of "Mrs. John Jones" is very destructive of the overall goal of the project to provide unified family tree. In the past, entering a name in that format was used by members of the Church to do that Temple work without going to the effort of identifying the wife. In some cases, especially at the end of lines, another generation of Temple work was done by identifying the parents of an ancestor as "Mr. Jones" and "Mrs. Jones." Often, this was done mechanically without any evidence whatsoever as to the identity of the parents of the last person in a line. I'm certain that I could go into Family Tree and find many, many similar entries.
I'm sure that the argument exists that these people need to have their Temple worked done and so what is the harm? In reality, this is one of the worst examples of researchers who are concerned merely with the names and dates of their ancestors and have no further interest at all in establishing anything concerning their lives. The interesting thing is that the Family Tree program will readily allow people to enter this minimal type of information and qualify the names for Temple work. Accomplish this all the user has to do is to put in the made-up name, and approximate birthday and enter the death date as "deceased." This of course does not avoid the moral dilemma of the individual submitting such information that the information is correct.
These examples do not even begin to address the real issue of what kinds of information need to be included in the Family Tree. Here is a screenshot of a portion of the pedigree presently showing for one of my remote Tanner ancestors:
The problem is that the individual marked with the arrow as number one is actually the end of the line and has been for some period of time. I might note, that a distant relative may have discovered actual documentary evidence extending the line into England, but for many years none of the family members have been able to find anyone past the original English immigrant, William Tanner. Now, it may well be that there was a "William Tanner" born in Rhode Island and 1687 with the father named William Tanner born in Great Coggerhall, England in 1660, but there has never been any documentation showing that fact. So all of the information in the large rectangle is entirely fictitious.Here is another screenshot showing the detail of the supposed Henry Tibbets ancestor:
This shows that he was born in Dover Neck, Strafford, New Hampshire in 1635. This is interesting and rather complicated. Let's just say that the historical figure of Henry Tibbets is well-documented as being born in England and living his entire life in New Hampshire. He was never married to anyone named Sarah Stanton. In fact the entries for Henry Tibbets include the following from another screenshot from Family Tree.
In addition, the entries for marriages for Henry Tibbets' father go on for an additional eight marriages. of course, There absolutely no sources attached to support any of this.
Unfortunately, this example which mostly illustrates the accretion of errors over 150 years is neither unique nor uncommon. In fact, it is only through watching certain key ancestors that additional information of this type is prevented from being added to even more recent ancestors.
The types of information that should be added to Family Tree does not include speculation and fiction.