Wikipedia:WikiProject Kentucky/GNIS cleanup

Per Project:Reliability of GNIS data there are many (sub)stub articles on Kentucky places that have been created sourced to GNIS entries. A lot of them falsely designate things as "unincorporated communities", which is several GNIS importers' catch-all equivalent to the the GNIS catch-all feature classification code "populated place". GNIS has many errors, and all of these (sub)stubs need cleanup.

Detailed problem statement
The GNIS computer database of names doesn't systematize Kentucky geography in the way that it actually systematizes. Wikipedia should be right and informative.

In many cases, even the one thing that the GNIS is still considered reliable for, co&ouml;rdinates, is not useful in Kentucky, because things actually moved around a lot. The GNIS co&ouml;rdinates are a fossilized snapshot off one set of mid-20th-century topographic maps at one point, and a certain professor took it upon himself to correct their many errors.

In places like California (c.f. Project:WikiProject California/GNIS cleanup task force) many GNIS "populated place" records are hot springs, or vanished ranches, or erstwhile stops on settler trails, or Southern Pacific railroad sidings. In places like Louisiana, the GNIS has got a whole load of errors such as designating bayous as "populated place" because they were on a USDA soil map and the GNIS collator in the 1970s was slipshod, and Louisiana genuinely has villages and plantations which should be called that. The way that the GNIS is wrong varies from state to state.

In the case of Kentucky specifically, most of these (sub)stubs are real locations, but more often than not they are not "communities" but old post offices or old mining towns or old freight railway stations that served the mines. (Unlike Louisiana, Kentucky has historically had a very flexible definition of "city" and tends not to have villages.)

Other problems include the GNIS getting things arse-backwards. The GNIS database compilers, when they didn't know the name of something, tried to deduce it from the names of things around it, and only consulted maps, maps from the middle 20th century. There are authoritative sources from the Kentucky Geological Survey that give the right and local to the state names for things that got completely overlooked. The 20th century USGS maps knew about "Nigger Branch" on Big Creek for example, but (ironically, given the mid-20th-century push to get this word out of U.S. place names) didn't know that the KGS had the far less outright ethnic slur name "Jenny Lick Branch" for it already there and ready in 1918 in its Fourth Series reports. ("Lick" names usually come from a salt lick.) To this day, the USGS is still sticking to "Negro Branch".

Plan
The plan is to cover Kentucky using its Fork/Creek/Branch system, using a top-down approach starting from the Kentucky River and building out articles if there is enough in the sources for a major tributary subsystem starting with a Fork/Creek/Branch.

The goal is not a single "unincorporated community" in Kentucky. Kentucky doesn't in reality have these, and this is a lazy cop-out by both the compilers of the GNIS and the Wikipedia GNIS database importers who copied the GNIS data into Wikipedia as (sub)stubs.

Rationale
Simply put:
 * 1) Kentucky's primary geographical features are its creeks and forks.
 * 2) Sources document Kentucky by its creeks and forks, and so should we.
 * 3) Every article by dint of construction gets at least two good sources this way.

The human geography of Kentucky often follows its physical geography, the Fork/Creek/Branch system. Many communities are strip development along the routes of river tributary systems. At the end of World War 1, Kentucky began a mining boom, and many places on maps come from that time.

Although known for its family feuds over the course of its history, even Kentucky's families followed the Fork/Creek/Branch system, family members often having mines and post offices and stores close by one another, and even on the Forks/Creeks/Branches that bear their family names.

Even in the age of the automobile, many roads follow the river tributary systems, even the major state highways. They sometimes follow what Geological Survey sources will document as "gaps" that join the headwaters of one creek to another, going up one creek and down another.

The the Robert M. Rennick Manuscript Collection contains (amongst other things) histories of the post offices and communities in Kentucky, and although Rennick proceeds county-by-county, inside each county he proceeds creek-by-fork, and often the forks that cross counties share information in his post office histories.

Examples:

The publications in the Kentucky Geological Survey, especially the Fourth Series under Joseph B. Hoeing, follow the Fork/Creek/Branch system systematically, navigating a specific tributary system in a linear fashion from creek mouth to its headwaters, going in depth-first fashion into its tributaries along the way. These actually prioritize the tributary system over the county bundaries, including multiple counties and only parts of counties as the river courses dictate.

Examples:

Taken together, even just these two (sets of) sources alone provide each major Fork/Creek/Branch tributary system with two sources, Rennick for the human geography including the post offices, some of the histories, and some of the people, and the Kentucky Geological Survey for the physical geography, the creek system co&ouml;rdinates, and the mines ("entries") and miners.

Examples
Already created Fork/Creek/Branch system articles following this pattern include:


 * Kentucky River
 * North Fork Kentucky River
 * Troublesome Creek (North Fork Kentucky River)
 * Lost Creek (Kentucky)
 * Buckhorn Creek (Kentucky)
 * Balls Fork
 * The Forks of Troublesome
 * Big Creek (Perry, Kentucky)
 * Lotts Creek (Perry, Kentucky)
 * Trace Fork
 * South Fork Kentucky River
 * Goose Creek (Oneida, Kentucky)
 * Little Goose Creek (Kentucky)
 * Horse Creek (Kentucky)
 * Laurel Creek (Kentucky)
 * Martins Creek (Kentucky)
 * Beech Creek (Clay County, Kentucky)
 * Red Bird River
 * Bear Creek (Red Bird River tributary)
 * Big Creek, Kentucky

You can see from Special:Whatlinkshere on several of these articles how they have very effectively vacuumed up a whole load of prior "unincorporated communities" (sub)stubs, which were sometimes sourced to Rennick's book on place names (because it was easier to find and use by people wanting to "save" their substubs than Rennick's actual far more detailed magnum opus: the entire set of county histories that the book just skims).

Just a very few examples: Grace, Kentucky

Hooker, Kentucky

Dwarf, Kentucky

Brightshade, Kentucky

Questions and answers

 * Can I help by nominating "unincorporated community" articles for deletion?: No. The end goal has these as redirects pointing to the Fork/Creek/Branch system articles.  We don't get there by clogging Project:Proposed deletion or Project:Articles for deletion, as it doesn't need the administrator deletion tool to just boldly make a redirect.  Plus, it takes quite a while to build a decent Fork/Creek/Branch article, often longer than the AFD one week deadlines allow.  Don't force the process with wasteful deletion nominations.  Write creek articles and do the redirects, instead.
 * So everything is a redirect, then?: Sadly, also no. On a very very rare occasion, Rennick has nothing.  Such was the case for  for example.  But notice that it was researched againsst Rennick very hard before coming to this conclusion.  You will almost never hit this scenario, given how in-depth Rennick is.  Another very very rare case is where Rennick disputes apocryphal places such as  for example.  A rare but slightly more common case is where the GNIS record is not in fact sourced to Rennick (as many are) in the first place, as was the case for.
 * But don't post offices grow communites? A post office existing implies a community, surely?: No.  Historically these are family houses, and ironically the communities that they served were the people who lived along the creeks that the post offices served.  Avawam post office that is now on Big Creek is an extreme example that, although you wouldn't know it from people who provide guides to where post offices are now, changed families, changed creeks, and even changed counties.  The Tina post office on Montgomery Creek off Troublesome was an Everidge family post office that didn't define a "community", because it too moved around a lot.  It went whereever the family (home) was. Furthermore, and amusingly, the Fork/Creek/Branch system has many examples of where the USPS did not allow people to name the post offices for their communities.  The USPS rejected three attempts to establish a Ball post office on Balls Fork. As icing on the cake we have examples such as the Lost Creek post office of Lost Creek whose enclosing village is actually called Troublesome, and the Rowdy post office on Troublesome Creek whose environs are actually called Stacy (with a local family by that name in creeks all around during the mining boom).
 * So it's just Rennick's post offices histories that we need?: No. Leaving aside that, as aforementioned, there's also the massive library of geological and scientific sources from the  Kentucky Geological Survey reports, Rennick has more than that.  We can give Rennick's annotated topographic maps for the quadrangles that each Fork/Creek/Branch system crosses.  And, ironically, the editors at  missed that Rennick had a whole book on the numbered place names in Kentucky .  Then, as Trace Fork and Troublesome Creek (North Fork Kentucky River) exemplify, once you know how the geography of Kentucky actually systematizes, books on the train lines that used to go along the creeks to the mining towns show up, as do hydrological reports about flood control from the U.S. Army Corps of Engineers, and all sorts of other stuff.