Jack, I think Elliot gave the best answer:
> On Mon Dec 01 2014 at 12:56:30 AM Elliott Plack <elliott.pl...@gmail.com> >> wrote: >> > > . My opinion is to use the full name with the post directional and let map >> data users (or humans) choose what to ignore. >> >> Here are some reasons Usually all areas in the county use the same system, with directions >> denoted NE, SE, NW and SW from some standard point, although some cities >> tend to ignore the suffixes. Also, signage is inconsistent--some street >> signs bear the suffix while others on the same street don't. >> >> I cannot always rely on signs. That short sign names may be from the cost of creating the sign or there's not enough room to display all the information. For example, many of the major intersections of Phoenix AZ have lit street signs at intersections. Compared to Tempe AZ or Avondale AZ which use these monster lit signs, the Phoenix signs only have enough room for the street name, say, Washington of E Washington St and the hundred block. To remove the E and the St from OSM data would not be a good choice. Other minor non-lit signs have the full name. > The majority of the Tiger data doesn't include the suffix. >> >> I cannot always rely on Tiger or the USPS in some cases. I seen the USPS address correction tools remove street types even if they are on the sign and the locals use the full street name. There are a number of these loop things in the Phoenix area.[0] Without the postdirection it is a mess trying to find the correct addresses. There are odd and even addresses on all sides of the loop. The Paradise Village Parkway is one of the easier loops. I cannot agree with Brian's statement. The FGDC doesn't make any reference to pre-directionals when used for a >> divided road (dual carriageway). The problem with these roads is addresses >> don't use the pre-directional, but road signs do. Here's one example: >> >> http://www.openstreetmap.org/way/295032159#map=16/41.0291/-73.7359 >> >> The address of the highlighted building is "333 Westchester Avenue". >> Along this divided street, the odd building numbers are on the south side >> and the even numbers are on the north. >> > I fly into this area. I've never been there before. How would this incomplete address help me if I have no bearing as to east or west or odd and even numbering schemes? Some of the roads have east some have west some have nothing. "Westchester Avenue West" would at least help find the street. > For divided highways, one may see a directional modifier on the signs, but >> on the map it's never included, >> > Yes but there's so much data that can be displayed. I'd beg for a butt ugly map that displays all the hidden data that is in OSM. There's other uses for the data besides map tiles. - House 333, Westchester Avenue East, White Plains, Westchester County, New York, 10577, United States of America <http://www.openstreetmap.org/way/295032159> - Wow! How did Nominatim find all the complete address information based off this incomplete address!? or if it is included, is done through super relations. I.e., I-84 contains >> two relations, one is I-84 Eastbound and the other I-84 Westbound. The >> contains way's whose role is east/west, but are named I-84. >> >> For divided roads, it seems it's best to put the directional modifier in >> the relation's role and omit it from the name. >> > Oh please don't do that. [1] Why does OSM Inspector fail trying to find an address and road match in this area? Notice the long black lines away from the obvious road right next to a house. I don't see the same problems with addresses and roads with the full street name such as "East 241st Street". The OSM Inspector address tool [2] pointed out some major issues in the Mesa AZ area after it was released. Someone decided to remove all the east and west prefixes on the streets. The change comments base this removal on some wiki page section that no longer exists. The unfortunate result with this choice is that there are several address range overlays going on in this area. [2] and [3] The really really nasty result of the two towns growing together, is that Tempe ends around 2600 ish East University Drive while Mesa starts around 2550 ish West University Drive. An address with only 2400 University Drive leaves me hopeless. I've cleaned up a bunch of issues in the area with the help of OSM Inspector. However, I feel like the street names are crap in the area because I don't know that I've fixed all the problems with the prefix data. There would be a similar problem if someone decided to remove all the suffix data on the loop streets. Finally, Elliot pointed out the interesting FGDC standards [1] that helped me to surf into [4], [5], [6] and [7] When you look at page 49 and 50 of presentation PDF [4], one of the goals is an open address exchange. The specs want you to remove abbreviations before sharing addresses. OSM data fails. If I just want the data in addr:* tags, then most of our information is incomplete. [6] and [7] show the broad impact of OSM. The Open Parcel data using the OSM editing tools is an interesting idea! [7] Regards, Greg [0] http://www.openstreetmap.org/#map=16/33.6006/-111.9833 [1] https://www.fgdc.gov/standards/projects/FGDC-standards-projects/street-address/06-09-21.ASWG.Presentation.URISA2006a.pdf [2] http://www.openstreetmap.org/#map=19/33.42215/-111.87739 [3] http://www.openstreetmap.org/#map=17/33.42281/-111.83178 [4] https://www.fgdc.gov/standards/projects/FGDC-standards-projects/street-address/06-09-17.ASWG.ExecSumm.Rev4a.pdf [5] https://www.fgdc.gov/schemas/address/addr_type.xsd [6] *Mr. Steve Coast* Telenav, Inc. https://www.fgdc.gov/membership/ngac/membership [7] http://openparceldata.wordpress.com/
_______________________________________________ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
