Proposed mass edit for street names in VIC: removal of spaces surrounding hyphens

Communities Oceania
, , ,
1

(this mass edit is not going to be made due to the em dashes issue raised below)

I would like to propose a mass edit that removes spaces surrounding hyphens in Victoria. Most of these road names appear to be added by lots of mappers using the Vicmap Transport layer, which displays these road names with a fair bit of space around the hyphens (image below). In the original dataset, there is not a single road name that contains <space>-<space>, but there are 5811 name and 2975 alt_name tags that include spaces around the hyphens. This mass edit would remove these extra spaces from 5088 and 2971 of them.

image
image313×351 24.9 KB

This is a problem because road names should exactly match addr:street. There are no addresses in VIC’s database that have spaces around the hyphens. Also, it is often not possible to rely on street signs/“what’s on the ground” for these hyphenated road names, as many only have signs that show destinations instead of road names - example.

I propose that these road names (both name and alt_name) are edited mechanically, replacing " - " with "-":

Myrtleford - Yackandandah Road -> Myrtleford-Yackandandah Road
Healesville - Koo Wee Rup Road -> Healesville-Koo Wee Rup Road

I will use the below Overpass query, which ignores various cycling routes, tracks, highway exit ramps, and other informal names. The name tag on lines 3 and 4 can be changed to alt_name to match that tag instead.

type:way and
highway~"(living_street|motorway|motorway_link|primary|primary_link|residential|secondary|secondary_link|service|tertiary|tertiary_link|trunk|trunk_link|unclassified)" and
name~" - " and
name!~"([Rr]amp|Loop|Access|Gate)"
in Victoria

I’ve made a list of all road names that will be changed in this mass edit for easy checking:

I’ll leave voting open for two weeks, and if approval is above 90% and no major issues have been identified I will complete this edit using Level0 and my text editor, uploading with an account specifically for mass edits. I can filter lines to replace in my text editor, so any other tags that contain " - " will not be changed.

2

A word of warning…

Style guides, orthographic manuals, and typography, generally distinguish a hyphen (-) and a dash (–). Broadly speaking, a hyphen indicates (strong) conjunction between the words, as in surnames (Marion Jones-Thompson, Wilkes-Barre disease) and compounds (ex-wife, quasi-science). A dash indicates a connection between distinct things (Anglo–French relations, Paris–Roubaix race). Following the convention allows one to distinguish (hypothetical)

  • Lennard-Jones Law (named after John Lennard-Jones), and
  • Lennard–Jones Law (named after a certain Lennard and a certain Jones).

So, the “proper”, “careful”, “nerdy” way to spell this road name is actually with a dash: Myrtleford–Yackandandah Road.

Now, I’m not advocating that we introduce (or even allow) typographic dashes in OSM names. However, a common convention using “ASCII only” environment is to use a spaced hyphen instead of dash, to indicate a connection. That also allows you to finely handle cases including spaces, like

  • New South Wales - South Australia Highway from
  • New South Wales-South Australia Highway (there’s no such place as “Wales-South” )

(And, of course, changing typography and style from the source system during an import process is a can of worms in itself, since the addr:streets must be adapted as well).

So, I’ll refrain from voting, but just sayin’ that there’s a reason why those spaces have been introduced in the first place.

4 Likes
Multilingualism in Serbia
3

That’s an excellent point, and something I was thinking about too.

For example Relation: ‪Ku-ring-gai Chase National Park‬ (‪5989782‬) | OpenStreetMap “Ku-ring-gai Chase National Park” where “Ku-ring-gai” is a single word, whereas “Healesville-Koo Wee Rup Road” is a distinctly “Healesville” [to] “Koo Wee Rup” “Road”, so we should aim to distinguish these and I agree that forcefully including a space or using a dash (–) is a good idea.

4

That’s very interesting, I did not know that using spaces like that was a common pattern to replace dash characters. It looks like instead of this mass edit we may need to pick a style to use for consistency of these road names.

As some pretty rough numbers, 699/1517 (46%) of unique hyphenated VIC road names in OSM use spaces around the hyphens.

1 Like
5

Yes, provided that we stick to hyphens. En dashes will just invite ongoing issues for mappers.

6

There was some discussion on Discord in #oceania about this: Discord

I think en dashes are the most correct option, I agree it may cause some issues with mappers but:

  • popular editors auto-complete addr:street based on nearby streets, which would copy the en dash
  • given these road names are unique, it should be simple to automatically check for any inconsistent usage
7

Please don’t. It’s just pushing the responsibility onto the mapper. Not only being a PITA it also violates the on ground principle, as AU street signs do not have en dashes.

1 Like
8

What style do you prefer then?

I’m not suggesting these be used outside of Victoria - I know that at least Queensland is very inconsistent in using punctuation in these roads. All of Victoria’s destination-to-destination roads are consistent in data they publish.

Of the roads I checked in VIC, most don’t have a street sign. The few that do seem to be inconsistent with dashes and even new lines, with things like:

PlaceA-
PlaceB

PlaceA
- PlaceB

or no hyphen at all.

I don’t think the on the ground principle applies as clearly to these roads - they don’t have a name other than being between two places.

1 Like
9

<space><hyphen><space>

Indeed, but it will never be an en dash because the typeface used in AU does not have one.

10

This reminds me of a story…

In the part of the U.S. where I started mapping, “Terminus-Terminus Road” is one of the most common forms of a road name, usually naming two localities named after long-gone farms named after long-gone farmers. Another common form is “Person Road”, named after just one person.

Often there’s mass confusion over the road’s etymology. Is “Irwin Simpson Road” named after a Mr. Irwin Simpson? Or is “Irwin-Simpson Road” named after the double-barreled Irwin-Simpson family? Or did the road originally run between Irwin and Simpson before it was extended in both directions and urban sprawl erased both place names? Is “Stonelick Williams Corner Road” named for the corner of Stonelick and Williams, or the corner where Mr. Stonelick-Williams farmed, or does it connect Stonelick with Williams Corner?

On street signs, as well as in phone books, newspaper articles, road atlases, and GIS databases, the name would appear with a hyphen or maybe just a space. Like the postal service, none of these media recognizes any character beyond ASCII, and the hyphen is interchangeable with the space for all anyone cares. After all, you’re just as likely to see all punctuation stripped off and a space added after “Mc” or “O” to avoid confusion in all caps. Even when the road is obviously named after its endpoints, the signs can still disagree vigorously about the spelling and punctuation, as in Socialville(–)Foster(s) Road. If only there were a local convention to pad the hyphen with spaces!

Map of Mason and Deerfield Township, Ohio, showing 14 nodes in four colors by name.
Map of Mason and Deerfield Township, Ohio, showing 14 nodes in four colors by name.2646×1832 1.12 MB

Yet if you happen to find a reference to the road in a professionally typeset book, it very well may use an en dash, as long as the editor can figure out the etymology. Back in 2009, when I was a wee mapper, I thought it would be pretty cool if OSM would help data consumers achieve the same level of sophistication as those books. So I went around several counties in Ohio, replacing the hyphens with en dashes one by one as I determined that the names referred to termini. “Murdock-Goshen Road” became “Murdock–Goshen Road” – or so I thought.

Unfortunately, I was making these edits in Potlatch, completely unaware of a compatibility issue between Flash and macOS. I was accustomed to typing ⌥− on my U.S. English keyboard to insert an en dash. It caused Potlatch to display “Murdock-Goshen Road” as before and Osmarender and Mapnik to show “MurdockGoshen Road” in CamelCase, so I thought the software lacked adequate fonts to show the en dash. Turns out Flash actually bound that key combination to the soft hyphen: “Murdock­Goshen Road”. Even the website’s data browser was no help, because the tag value happened to line-wrap in a manner that exposed the hyphen. It took me eight months to notice it.

Won’t anyone please think of the Potlatch users out there! :sweat_smile:

3 Likes
11

Interestingly, AS/NZS 4819:2011 uses an en dash in these road names, calling them “Origin–Destination” roads. (the standard also says new road names like this shouldn’t be added, but these are clearly in use and/or were named before this standard)

4.4.9 Origin–Destination road name
A road name shall not consist of the conjoined names of places or localities found along or
at the ends of the road (e.g. ‘Newtown–Kelston Road’)

You can also see how inconsistent QLD is here, they’ve somehow managed a spaced en dash! https://www.tmr.qld.gov.au/business-industry/business-with-us/alliances/country-roads-connect-program

1 Like