Ha, I was still editing my OP before I realized you posted. I added something in there about the existing preferred fact before I saw what you wrote.
"But, that is your choice. You can make either fact preferred or alternative or discarded (with source kept). FTM just gives preference to a fact with a date over one without a date. If you want to keep the one without the date, just change discard to alternative."
You're right. I do have the choice to keep it, but it gets relegated to "alternate" status. That's my call to make. Why do I have to go through the steps to restore my preferred fact to preferred status after I merge this 1870 record?
My argument though is that the default choice encourages a bad practice and creates extra work for people who don't follow the bad practice.
"I don't know if I follow you exactly, but, again, that is your choice. "
I'll try to explain it better. Say I have two pre-existing birth facts, a preferred one that just says "Norway" and an alternate that says "abt. 1833 in Norway". The Web Merge Wizard wants to merge my preferred fact with the incoming fact. Why doesn't it suggest that I merge it with the other alternate fact? I don't get the reasoning behind setting up the default option here.
"FTM gives you that choice. It just offers to make some incoming fact preferred, but you can override. FTM will suggest a date preferred over no date, an exact date over a year only date, and so on. But, that is just a suggestion. You can discard or make alternative either one."
This is the part I added to the OP before I saw your response. It actually doesn't give me the choice to retain my preferred fact. If I keep it, it must become alternate.
My argument though is about why FTM is encouraging me to prefer one over the other. I say it's based on the false assumption that anything with a date is preferred over anything without. That's not always going to be the case. It's a false assumption that sets up the default options that encourage me to obfuscate my sources and risk having to redo work I've already done.
"I notice you have a person without a birth date in your file. This is something I NEVER do. I ALWAYS give a birth date for everyone in my file. There is ALWAYS a way to estimate it. People are too hard to find even when you do give birth dates to all - a relatively large database would be unmanageable without birth dates."
I add the facts as I come about them in the records. In this case, the first record was an indexed marriage record for a child that listed his father's birthplace. There isn't another record for him yet from which a birth year could be estimated.
"That said, the bigger problem to me is to not be able to merge an incoming fact with an alternative date. If I have an exact date of birth of say, 01/23/1901, and then a series of incoming census records of either "1900" or "1901", I'd like to be able to merge them in their little alternative group (ie all the 1900s merged, all the 1901s merged, etc). Now, I have to go back to the person tab for that person and right click > merge facts to merge those kinds of alternative facts into groups. Similarly, I may have Richard Carlson Smith as 2 R. C. Smiths, 3 Richard C Smiths. I'd like to be able to merge these alternative names together as they are merged than later in a separate step. But, that would increase complexity and FTM probably wants to keep thins as simple as possible."
I would argue that's a bad practice. For myself, I would never risk doing the same work twice by merging that information. Different sources say different things. Why no longer be able to see that at the fact level? If I were later to decide that one fact is true over the others, then I would create a new preferred fact to which I would copy and paste the non-conflicting sources.
I agree about the myopic focus on the preferred fact. I add everything as an alternate unless it's an exact match. For me, R. C. Smith ≠ Richard Carlson Smith ≠ Richard C. Smith. I would never merge them. (You, of course, can do whatever you like, even if, from my ivory tower, I think it's a bad idea.) This means that I end up with lots of alternate facts, so whenever I add anything through the Web Merge Wizard or through Ancestry.com, I almost always have to do cleanup. It's a PITA. We should discuss this more, identify some least common denominators, and submit independent enhancement requests.