[Talk-us] US Interstate exit junction exit_to tag

[Alan Mintz]

At 2011-04-07 01:31, Mike N wrote:
>On 4/7/2011 12:53 AM, Alan Mintz wrote:
>>At 2011-04-06 15:26, Nathan Edgars II wrote:
>>>On 4/6/2011 5:59 PM, Alan Mintz wrote:
>>>>At 2011-03-28 12:40, Ian Dees wrote:
>>>>>With that in mind I think it's important that the exit_to tag only
>>>>>include verbage on the sign (and not stuff we make up).
>>>>
>>>>IMO, not all the verbage on the sign. I've been tagging name as the name
>>>>of the exit according to the relevant authority (e.g. CalTrans in CA). I
>>>>accept that people now want to change this to use exit_to instead of
>>>>name. However, some exit signs also have destination information, like
>>>>the name of the city or local tourist attraction. This secondary
>>>>information I've been putting in a towards tag. I would also like to
>>>>separate the name into a root and a directional component. e.g.
>>>[snip]
>>>This seems overly complex. The reason for including the text is so
>>>routers can tell the driver to take the exit marked [foo]. What
>>>benefit is there in separating street names/route numbers from other
>>>destinations?
>>
>>Routers and renderers are not the only consumers. I'm trying to get the
>>data modeled correctly. Logically, the name is the important part. If
>>necessary, for reasons of speed or space, the towards part can be
>>eliminated. If exit numbers are unimportant to a particular consumer,
>>they can be eliminated to. That is the point of breaking it up into
>>multiple fields. As far as complexity, it seems quite simple to speak
>>"take " + iif($ref != "", "exit " + $ref + ", ", "") + $exit_to +
>>iif($towards != ""," towards " + $towards, "").
>
>   It's one thing to model the data correctly, but for a crowdsourced 
> project like OSM, how do you translate that to thousands of people 
> knowing how to correctly enter their local exit information?   I don't 
> work for the DOT, and would have no idea how to enter the local exit 
> information according to that model.   Who decides what 'towards'
>is to be, and how is that clearer than just trying to match a sign?

I don't think you have to work for the DOT to know, and I don't think it's 
complex. Certainly no more than picking from among a dozen kinds of ground 
cover or barrier types, assigned names that are not widely known in the US. 
Or water feature names that you'd have to be a civil engineer or geologist 
to distinguish among. Worst case, they just dump the entire sign in exit_to.

Case 1. The exit_to (and exit_to_dir) are the street name at which the ramp 
terminates in 99% of the cases. e.g.:

"Exit 239 / Zzyzx Road"
is tagged
ref="239" +
exit_to="Zzyzx Road"

"Exit 61 / SR-79 North"
is tagged
ref="61" +
exit_to="CA-79" +
exit_to_dir="North"


Case 2. Very occasionally, there will be more than one street name shown, 
usually when the ramp ends at or near a point where a street changes name. 
Use semicolons to place multiple values in the exit_to and exit_to_dir 
tags. e.g.:

"Exit 73 / Diamond Drive / Railroad Canyon Road"
is tagged
ref="73" +
exit_to="Diamond Drive;Railroad Canyon Road"

"Exit 183 / SR-247 South / Barstow Road"
is tagged
ref="183" +
exit_to="CA-247;Barstow Road" +
exit_dir="South;"


Case 3. Even less frequently - more in rural areas where there is a single 
offramp for a small town - either a town name or attraction follows a 
street name, or there is just the town name or attraction on the sign. The 
town name or attraction goes in the towards tag. This case is often used 
for interchange ramps with another motorway, too. e.g.:

"Exit 3A / I-15 North / Sacramento"
is tagged
ref="3A"
exit_to="I-15"
exit_to_dir="North"
towards="Sacramento"

"Exit 139 / Keene"
is tagged
ref="139" +
towards="Keene"


>    You have made a good point that the way locals refer to an exit could 
> be valid as the name= tag.    If you have knowledge of the local 
> designation of all exits you've tagged, it might be easier to ask for a 
> rollback of just the edits where you are the author.

I have knowledge that those are the names of the exits based on the 
signage, and driving over wide areas and interfacing with others in 
delivery and other businesses, EMS, etc. for over 30 years. My experience 
in 20 years of business, traveling throughout both rural and metropolitan 
areas of the US is the same.

Does anyone have examples of places where my suggested model does not work?

BTW, my intent is to import, once I'm done tagging from survey pictures, 
applying the additional milepost and postmile data from the state to 
existing junctions, and adding names of those that I haven't 
visited/tagged. This should be fairly straightforward and high quality, 
filling what may be a significant gap quickly.

--
Alan Mintz <Alan_Mintz+OSM at Earthlink.net>