Why Most American Street Names Should Be Abbreviated in OpenStreetMap

Anatomy of a Street Address

Before I get to my main points of argument, I'd like to break down the two most common types of street addresses used in the United States and give names to their different parts.  There are some rare exceptions, but nearly all of the address systems I've seen fit one of these patterns.  Usually, a single address system covers an entire county.

Optional Directional Prefix Style

The majority of US address systems seem to use this pattern.  It works equally well in grid-like and non-grid-like street patterns.

4242 S Champion Ave E

This is the house number.  It's obviously not part of the street name.

This is the directional prefix, which isn't always present.  If a north–south street crosses the east–west axis of the address grid — or if an east–west street crosses the north–south axis — then a directional prefix is attached to the street name to distinguish between different halves of the same road.  *

This is the core of the street name.  Every street name has a core.  Sometimes, a directional word like West might actually be part of the core, sometimes to differentiate between different sides of a circular or semicircular road.  The core should not be abbreviated.

This is the street type suffix.  While somewhat arbitrary and certainly not a foolproof way to give a road a useful classification, it can hint at the road's age.  Some street names don't have a street type suffix.

This is the directional suffix.  When there are two parallel streets with the same name, sometimes on opposite sides of the city, the directional suffix is used to distinguish between them.  The directional suffix is also used sometimes to distinguish between different sides of a circular or semicircular road.

*Personally, I've always considered the directional prefix to be the sign of the house number (positive or negative) and not actually part of the street name, but that's another argument altogether.

Quadrant Suffix Style

This is another very common address pattern.  It works best in address systems with very grid-like streets, but I've seen it used in less-orderly places, such as Fairfield County, Ohio.

1337 Rainbow Dr SW

This is the house number.  It's obviously not part of the street name.

This is the core of the street name.  Every street name has a core.  Sometimes, a directional word like West might actually be part of the core, sometimes to differentiate between different sides of a circular or semicircular road.  The core should not be abbreviated.

This is the street type suffix.  While somewhat arbitrary and certainly not a foolproof way to give a road a useful classification, it can hint at the road's age.  Some street names don't have a street type suffix.

This is the quadrant suffix.  It makes the street name specific to one quadrant of the address grid.  In quadrant-style address grids, every street has a quadrant suffix.

Agreement with Signage

On official road signs, it is almost universal practice to abbreviate the prefixes and suffixes of street names.  The core of a street name is seldom abbreviated, even if it contains a directional word or a word that looks like a street type suffix.  This practice reduces confusion which might otherwise arise, as covered in the Unambiguity section below.

There are some rare districts and municipalities that write out the full prefixes and suffixes of a street name on the signs.  The motivation for doing this is probably a desire to look cultured.  Typically, these signs also use fancy fonts that are not approved in the MUTCD

Official References (Examples)

In this section I'll cite official government use of street names, and show that they use abbreviations just like street signs.  There may be occasional inconsistencies, such as "Av" instead of "Ave", or putting periods on abbreviations; remember, the government is made up of imperfect people.

Franklin County Auditor

The Franklin County Auditor's website has a property search feature.  When searching by address or intersection, the site advises users to omit the street type suffix from the entered search terms, to improve the probability of a successful search.  This is probably to deter people from entering a fully expanded (and, in my opinion, wrong) street name such as "Broad Street", which would not return any results.  An example search result lists the property's address as "4494 W Broad St" and its legal description starts with "Broad St". 

Ohio Department of Transportation

I'll concede that most of the references to street names on ODOT's website use expanded street type suffixes.  But those pages are written with the intention of user-friendliness to the average motorist.  Sometimes they even refer to US routes as State routes.  (I have to wonder, though: wouldn't it be more user-friendly to refer to streets the way they're labeled on the signs?)  But in their internal documents, such as the Straight Line Diagrams, abbreviated street type suffixes are consistently used, even where there is ample room to write out the fully-expanded name.

United States Postal Service

The United States Postal Service is rather particular about address formatting.  They have a list of standard street suffix abbreviations they'd like people to use.  While this addressing guide doesn't assert that this format is the preferred way to write a street name in all cases, it sets a strong standard for writing addresses which has implications for geocoding, discussed below.

Unambiguity

There are some street names which contain, in the core part, words which look like prefixes or suffixes.  If all the prefixes and suffixes were to be expanded, it would become difficult to distinguish the prefixes and suffixes from the core of the name.  Then, if a renderer or routing program automatically abbreviates the street name when displaying it to the user — as the OSM wiki suggests — then the result might be an entirely different street name.  Some examples follow. 

Prefix Core Type D/Q Suf. Location
  North Star Rd   Columbus, OH
W North Broadway     Columbus, OH
E North Broadway     Columbus, OH
  Broadway     Grove City, OH
  Park St   Columbus, OH
  Westpark St   Columbus, OH
  West Case Rd   Columbus, OH
  Case Rd   Columbus, OH
  Lane Rd   Upper Arlington, OH
W Lane Ave   Upper Arlington–Columbus, OH
E Lane Ave   Columbus, OH
  Parkway Ln   Hilliard, OH
  Northwest Blvd   Upper Arlington–Grandview Heights, OH
  North St   Columbus, OH
W South Ave   Plain City, OH
E South Ave   Plain City, OH
  West Ave   Plain City, OH
E West Ave   Plain City, OH

There is no automatic abbreviation algorithm that can, given the fully-expanded forms, reduce all of these street names to their correct abbreviated forms.

Geocoding

When people search for a US address, they're most likely going to type in the address as it would appear on a piece of US mail.  If all of the US roads in OpenStreetMap have their name changed to the fully expanded form, a geocoding program is going to have trouble finding the correct street.  All of the address tagging will have to have an addr:street value that's different from the name value of the associated street, which will irk those people who value topological information storage.  (The only time that should really be necessary is where, for example, people living on Old Walker Rd still have Walker Rd addresses.)

Summary

As far as I'm concerned, the "abbreviated" form of a US street name as it appears on signs is that street's official, full name.  There is no advantage storing the expanded, or "oral" form of the street name in the OSM data, because it rarely appears written that way.  Just because street names aren't often abbreviated in the UK doesn't mean that a street's actual name can't contain abbreviations.

After elaborating on my points of arguments, I now realize that my position is not as strong as I originally thought.  Still, I stand by my opinion.  A road's true name in the US often contains something like Rd or Ave, and the OSM data should reflect that.