The current policy is one of those things that's easy to say, but nearly impossible to actually do if one tries to apply it to all situations. I fear it's naïve.
I'm certain most people think it means "trimming 'USA' off the end pre-1776 makes me feel clever". Heck, I do that and I feel smart when I do it. That's a far cry from doing the research to figure out what the place was really called at the time. Just a quick glance at the lists of names and dates and maps for all the New England colonies makes it obvious how unreasonable it would be to actually do what this policy actually says, and how confusing the result would be, and how unhelpful the average WikiTreer would find the resulting information to be.
What does "at the time" even mean, when borders and jurisdictions were shifting throughout a given person's lifetime? I have an ancestor who, on his US Federal Census records throughout the 1800s, variously reported his birthplace as "Württemburg", "Preußen", and "Germany". This isn't because he got confused about where he was born. Those terms aren't synonymous or interchangeable. It isn't because his birthplace moved around. Each of them was correct at the time it was used. I'm literally imagining him and the census-taker needing to have a conversation about it every ten years, maybe picking up the newspaper to see who conquered Schale most recently...
We, especially-but-not-limited-to WikiTree leaders, should have a serious discussion about three questions:
1. What is the genealogical purpose of location information? What policy would best suit this purpose?
Both the at-the-time location and the current-day location are needed to help modern researchers locate records. When a location changed names several times, all of the variants are needed. Best for this is the free-text biography, to record all the pertinent information about places and how to find them.
2. What is the WikiTree purpose of placing location in a discrete data field?
G2Gers have noted that WikiTree provides a mapping utility keyed off of the location field, and they've pointed out that the link is useless when the place name in the field is archaic. Location appears in Watchlists. What else is it used for? It isn't validated or indexed or used for bulk operations, so I'm guessing there's no communal technical dependency happening.
This is important, and it is a completely separate question from #1. Which location information should be made available to WikiTree visitors, casual genealogists, serious genealogists? All of it, as much as possible. But which one of those potentially many variations should "win" a place in the location box? Seems like we could use the tech and enable some good new capabilities and knowledge by using one that maps to a physical location in some way, which means probably not archaic names that mapping tools can't find.
3. Are we serious about being an international site with a world tree, or are we content to be Europe (mostly Britain) and North America (mostly USA and Canada) centric?
A G2Ger made the excellent observation that removing "USA" from fields assumes that everyone on WikiTree knows even which continent "Maine" or "Plymouth" is on. How many of us can find Bihar on a map? Transvaal? Do we know what they were called 50 years ago? 500? Are we going to add support for someone to put "甘肃" in the location field? Are there even fonts for the extinct scripts some of our ancient nobles wrote? What about place names in non-written languages? Are we going to document that the Mayflower shallop went ashore at Meeshawn? It sure as heck wasn't "Provincetown Harbor" at the time they reconnoitered there. (That name was hard to find, by the way!)
I don't think the policy as stated is realistic for the scope of tree we say we're trying to create.
A more flexible policy - something like "do the best you can to achieve x goal in the location field, and describe variations in the biography if those distinctions are useful" - seems like it would better adapt to history, geography, language, experience level, etc. at scale.
Flexibility is also more consistent with our actual practice for LNABs. Given that the spelling of older LNABs varied considerably during a person's lifetime and there often was no single right spelling, we do the best we can to pick one for the LNAB field, in whatever way enhances findability and understanding, and we put the variations in OLN and/or the biography.