Get a sneak peek at whats next for Permanent Hazards on our April 7th Office Hours!
Post by qwaletee
Is there a public version of that form?
qwaletee
EmeritusChamps
EmeritusChamps
Posts: 2939
Has thanked: 188 times
Been thanked: 958 times
Send a message
US Champ / Country Manager | State Manager NY, NJ, PA, CT, MA, RI, VT, ME, NH | Northeast ARC | Mentor | Responding to Map Issues

Post by qwaletee
I leave it in your able hands :)
qwaletee
EmeritusChamps
EmeritusChamps
Posts: 2939
Has thanked: 188 times
Been thanked: 958 times
Send a message
US Champ / Country Manager | State Manager NY, NJ, PA, CT, MA, RI, VT, ME, NH | Northeast ARC | Mentor | Responding to Map Issues

Post by qwaletee
I still have issues with CDP since they are arbitrary tracts designed for census taking and often do not reflect local community boundaries. I'd like to leave it as "differentiator of last resort," or to bolster a postal name above an incorporated name.

My order of preference:

1) (rare) Local name when universally used and well-researched
2) (rare) Postal name where it is corroborated by local usage or CDP, but only so long as both have similar boundaries
3) Incorporated name (unless it creates ambiguity or has a very strong, well-researched local objection)
4) Postal name, uncorroborated
5) CDP name
qwaletee
EmeritusChamps
EmeritusChamps
Posts: 2939
Has thanked: 188 times
Been thanked: 958 times
Send a message
US Champ / Country Manager | State Manager NY, NJ, PA, CT, MA, RI, VT, ME, NH | Northeast ARC | Mentor | Responding to Map Issues

Post by qwaletee
I'm OK with adding secondaries, but let's watch out for any naming conflicts that may create. I'm sure there's some some town with no dedicated post office where the town name matches a post office elsewhere in the state.

For secondary names, we should only encourage it where there is genuine usage of the alternate name, or common confusion, e.g., at boundaries. I wouldn't smack down an editor who gets overzealous, though. These are the sort of arcane rules that are difficult to make sharp and ready.
qwaletee
EmeritusChamps
EmeritusChamps
Posts: 2939
Has thanked: 188 times
Been thanked: 958 times
Send a message
US Champ / Country Manager | State Manager NY, NJ, PA, CT, MA, RI, VT, ME, NH | Northeast ARC | Mentor | Responding to Map Issues

Post by qwaletee
George, forest and trees. I don't care about the polygon for our map. I care about accurate naming that represents what people will use in the real world. That comes down to local usage, which is an impractical standard, because 1) impractical to research generally (there is no database of local usage), 2) locals are not always consistent with each other, 3) there may be multiple accepted local usages.

All the other sources - USPS, USCB/CDP, and municipal - are generally easy to research and consistent for a given time period. That's a plus for us. But they're driven by unique needs of their system, which sometimes skews them away from what people use. However, they CAN provide a simulacrum of local usage, especially as all four tend to converge, if imperfectly. And, of the three, where there are deviations form local usage, the municipal name is least likely to be jarring.

That's why my preference list relies on a combination of "consensus checking" and weighting bias towards municipal for both practical research and "least harmful" error.

So, like I said, I don''t care about our map polygon. I do care about figuring out a way to assign a name to any segment or venue that will likely be the same as what a local will use. I.e., when I hit that restaurant's web site, I want to plug the address they list into Waze, have Waze recognize it, and return the listing with the same name I entered (or one that's at least cognate).

The ways things get skewed:

Local usage: As mentioned, there is no set criteria for this, and no easy way to conduct primary research. Names are fairly stable, exact boundaries, where they even exist, less so. This identification may be influenced (to change) by the other sources listed here, as well as other factors that ca't easily be tracked.

CDP - primary purpose, historically, and continuing mostly today, is to create a boundary for reporting information to the local level for "unincorporated places." (New Jersey has no place that is unincorporated, but alternate names, such as a strong neighborhood identities within a city, can be treated as unincorporated place names.) They are partially driven by the need to define a renable local census tract. They TRY to have this correspond to a primary local usage, using real local names and real local boundaries (though often estimated or fudged a bit). But they will often enough be forced to use a non-predominant name, or skew the borders to meeet their tract requirements. There are many changes to the list for each decenial census, including splits, merges, boundary adjustments to meet the rules, even though there has been no underlying change in locality naming for any other usage (local usage, municipal, postal). Occasionally, their naming and boundaries become a self fulfilling prophecy.

USPS - has two different systems. Historically, the post office WAS the address, so you addressed letters by the post office name. If it didn't match the city, someone would try to figure it out. And names were not always unique. Over time, a requirement for uniqueness in each state developed, while at the same time, direct delivery using local names (municipal or otherwise!) was allowed by mapping city or other local usage to a station. The post office policy is now to prefer OFFICIAL STATE names (generally, municipality), except 1) not to change legacy names for the station name, 2) in some cases, a second office in the same official name boundary requires development of a non-official name, and 3) delivery flexibility should be allowed, but the USPS maintains an official list of local names (separate from station names), with only one name for any point in the map listed as preferred. The upshot is that if we would follow USPS, we'd have to use their city list filtered for preferred names only. They have definitely done a loy of the same research we are trying to do, but actually down to the local level... and yet, because of requirements for uniqueness, the preferred name may be different than normal usage, and the majority of mail in a place may get delivered using the non-preferred name. Hence, zip codes, which avoid the name altogether, and can subdivide or combine without local issues. The post office station name or preferred name rarely effects local change (i.e., no self-fulfilling prophecy).

Municipalities are very clear cut, with official surveys precisely defining the geographic boundaries, and most people will either use their city as a primary location name, or "recognize" their municipality for addressing. Of course, there are plenty of exceptional cases, including a relative few where the city name is hardly used at all, and quite a few where the municipal name is recognized but would not be given out as first choice. (Technically, that's not an exception to the first sentence.) Municipal name and boundaries are least likely to change to accomodate on-the-ground changes in local usage - the USCB gets to redo its CDPs every 10 years, and may take into acocunt local naming changes, and the post office will add new names as non-preferred names, and may eventually recognize that the preferred name has become subsumed by one of the other usages, and select the new practical preferred name as preferred in the state-city file. Nevertheless, it is rare for t he municipal name to seem strange to locals.

Interesting notes on CDP: When the UCSB requested comments for proposed rule changed for CDPs for the last census, it got only 10. Two of those were from New Jerseyans who were unhappy that the new rules would eliminate their CDP. To my knowledge, neither of their authors is likely to read this post. The Census Bureau thought the comments might have been due to mistakes in understanding the rule change, but elected to punt the whole thing and not change rules. CDPs are generally divided by some visual boundary, e.g., topographies, highways, which align somewhat with local perception of community boundaries. CDPs cannot be contained within or ovelap incorporated areas. That should exclude CDPs from NJ, since every area is incorporated. It isn't clear but apparently they only partially recognize "townships," while fully recognizing "towns," "cities," "boroughs," and "villages."
qwaletee
EmeritusChamps
EmeritusChamps
Posts: 2939
Has thanked: 188 times
Been thanked: 958 times
Send a message
US Champ / Country Manager | State Manager NY, NJ, PA, CT, MA, RI, VT, ME, NH | Northeast ARC | Mentor | Responding to Map Issues

Post by qwaletee
They have to be correct almost every time for the user. Develop a rule that does that, I'm on board.
qwaletee
EmeritusChamps
EmeritusChamps
Posts: 2939
Has thanked: 188 times
Been thanked: 958 times
Send a message
US Champ / Country Manager | State Manager NY, NJ, PA, CT, MA, RI, VT, ME, NH | Northeast ARC | Mentor | Responding to Map Issues

Post by unlimited1808
voludu2 wrote:Two Examples in SE PA. Is there anything like this in NJ?

Wayne, PA -- a postal name that overlaps portions of 3 different townships in 3 different counties, but is not associated with any municipality or CDP. When I lived there, But everyone locally knows where Wayne is. It's the whole area with a Wayne mailing address. Or at least the slightly higher-density area near the main line and the Wayne Hotel. It's a dot on the AAA map, but not an area with a defined boundary, unless you count the extent defined by the USPS Wayne postal routes.

Glen Mills, PA -- a postal name that matches no CDP or municipality, which overlaps 7 municipalities in 2 different counties. I've removed it from the map a couple of times. Boundary definition has a similar issue to Wayne. And the Glen Mills themselves were torn down long ago.
There's a lot of these situations in SE PA. Add Ardmore and Haverford....almost every common name along the Main Line is not a muni or a CDP.


Also, I don't ever recall the absolute rule of CDP first and foremost in PA. Many CDPs are uncommonly used and I'd argue no advantage to Waze. I can't think of any specific examples, but I know we've removed a couple from the map recently that were useless.
unlimited1808
Posts: 295
Has thanked: 63 times
Been thanked: 83 times
Send a message

Post by voludu2
I think that is the only way the current cities model can actually be used.

Waze will have to change this in the future, but the future is not now.
Putting useful place names on the map, by whatever contorted means, is the best use we can make of the current cities layer.
voludu2
Posts: 3098
Has thanked: 559 times
Been thanked: 863 times
Send a message

Post by voludu2
I don't think there currently is any proper facility in Waze to handle city names.
And I agree that massive renaming would not make sense.

For example: a person might search for 403 Mountainview Dr, Chesterbrook, PA, which is well-understood to people, but just a large development, not a post office, municipality, or CDP. The correct municipality for that address is Tredyffrin Township, but the post office is Wayne, which is in a different county and is not a municipality or CDP, and so should not be mapped at all according to the PA state guidelines. It wouldn't make sense to develop alt-city guidelines because we don't have definitive answers from Waze on how alt-city figures into address searching now or in the future, so any effort to add alt names for the purposes of address searching could turn out to be a complete waste of time

On the other hand, as jondrush suggested, It might make sense to selectively add "map places" like Wayne to the map by using that city name as primary on a small number of segments at that map point place, because it can help wazers orient themselves, just as those labelled points on a paper map help people orient themselves.
voludu2
Posts: 3098
Has thanked: 559 times
Been thanked: 863 times
Send a message

Post by voludu2
qwaletee wrote: 2) (rare) Postal name where it is corroborated by local usage or CDP, but only so long as both have similar boundaries
Color emphasis mine.

There is an inconsistency in this rule.
If the postal name is corroborated by local usage, but not by CDP, then there is nothing for the postal name area to have similar boundaries with. Postal names will almost always be corroborated by local usage, because people will be familiar with the their own mailing addresses and the mailing addresses of businesses, so this is not a rare situation.

Two Examples in SE PA. Is there anything like this in NJ?

Wayne, PA -- a postal name that overlaps portions of 3 different townships in 3 different counties, but is not associated with any municipality or CDP. But everyone locally knows where Wayne is. It's the whole area with a Wayne mailing address. Or at least the slightly higher-density area near the main line and the Wayne Hotel. It's a dot on the AAA map, but not an area with a defined boundary, unless you count the extent defined by the USPS Wayne postal routes.

Glen Mills, PA -- a postal name that matches no CDP or municipality, which overlaps 7 municipalities in 2 different counties. I've removed it from the map a couple of times. Boundary definition has a similar issue to Wayne. And the Glen Mills themselves were torn down long ago.
voludu2
Posts: 3098
Has thanked: 559 times
Been thanked: 863 times
Send a message
Last edited by voludu2 on Tue Nov 13, 2018 10:48 pm, edited 1 time in total.