[Talk-us] Specifications (was: Bay Area trailer parks: "hamlet" ? Also neighborhoods & cities)
Michal Migurski
mike at stamen.com
Sat Nov 29 05:39:02 GMT 2008
Thanks Alan,
There's a bunch to chew on there, it's going to take me a few days to
move through the docs. It sounds like what you're suggesting, though,
is that the place=hamlet tag really is appropriate for neighborhoods,
which is fine. Are you commenting at all on its appropriateness for
trailer parks and apartment developments? I may convert some of the
Oakland hamlets to named landuse=residential to see what happens.
I don't actually know what the main OSM tile rendering will do with
these. =)
-mike.
On Nov 28, 2008, at 1:10 PM, Alan Brown wrote:
> In commercial data, a category called hamlet is use dfor either
> neighborhoods, or small unincorporated settlements. I've used them
> for neighborhoods in OSM mostly because I want to see them rendered
> - while it would be better to model them with descriptive names.
> (Formal point-of-interest - "city centers" - all fall into the same
> category, but have different "population classes". It's nearly
> identical to the module used for "place" point features, although
> population ranges are formally set for each place type. There's
> also ways to flag is something is a national capitol, state capitol,
> or county seat.)
>
> A lot of the discussion held here came up years ago when they
> designed the GDF specification - (Geographic Data Files). This is
> an exchange format. A lot of that specification has to with the
> format of the ascii file - which is irrelevant to our discussion -
> but it also deals with features types and categories. This is the
> format that various navigation systems and online map companies
> receive before converting it into their own proprietary access format.
>
> Here's a short wikipedia summary:
>
> http://en.wikipedia.org/wiki/Geographic_Data_Files
>
> Here's a link to an old version of the specification:
>
> http://www.ertico.com/en/links/links/gdf_-_geographic_data_files.htm
>
> This version (3.0) came out 13 years ago. It's since been
> significantly enhanced - both by ISO, and by Tele Atlas and NAVTEQ.
> The enhancements add a lot of subtlety, particularly to POI
> categories. The unfortunate thing is that Tele Atlas and NAVTEQ
> versions are not 100% compatible with each other, and no one fully
> complies with the ISO standard. The attributes of the 3.0 spec are
> largely followed.
>
> The Annexes are probably the best place to start, and it contains an
> appendix of feature categories - while chapters 5, 6, and 7 describe
> features in detail.
>
> While this spec doesn't resolve the "hamlet vs. neighborhood" issue
> - and it doesn't flesh out details for bicycle routes and paths - it
> has a lot of sophistication for things like different types of
> highway ramps, turn restrictions, dealing with road connectivity,
> etc., etc. I know they want to give people the ability to tag
> things in new ways, but if OSM has aspirations of becoming a
> routable database, things have to fall into predictable categories
> for software to interpret them.
>
> -Alan
>
>
>
>
>
>
> From: Scott Atwood <scott.roy.atwood at gmail.com>
> To: Michal Migurski <mike at stamen.com>
> Cc: talk-us at openstreetmap.org
> Sent: Friday, November 28, 2008 12:25:38 PM
> Subject: Re: [Talk-us] Bay Area trailer parks: "hamlet" ? Also
> neighborhoods & cities
>
> On Fri, Nov 28, 2008 at 12:13 PM, Michal Migurski <mike at stamen.com>
> wrote:
> Hi,
>
> There are a large number of mobile home / trailer parks mapped in San
> Jose, Santa Clara, and other parts of the South Bay. They're tagged
> place=hamlet, and I'm wondering if there's a better way to identify
> them? Beej71, if you're on this list I think a lot of these came from
> you.
>
> Examples here:
> http://www.openstreetmap.org/?lat=37.3995&lon=-122.01521&zoom=15&layers=B000FTFT
> http://www.openstreetmap.org/?lat=37.37096&lon=-121.89402&zoom=16&layers=B000FTF
>
> "Hamlet" is supposed to be "defined by national/state/provincial
> government" yet these aren't really defined by anyone except their
> owners. I personally view them as generally equivalent to named
> apartment complexes, and therefore not a place. May I suggest that
> they be redrawn as landuse=residential areas, with names defined?
>
> On a related topic, I'm also wondering how to handle parts of cities
> that are "places" or neighborhoods yet not administratively distinct,
> e.g. "West Oakland" (Oakland), "The Mission" (SF), etc. Would
> place=neighborhood make sense here?
>
> This part of West Oakland with two named apartment complexes (all
> place=hamlet) illustrates what I mean:
> http://www.openstreetmap.org/?lat=37.8096&lon=-122.29504&zoom=16&layers=B000FTF
>
> I think people (including me) have been using/misusing place=hamlet
> for places like neighborboods which have variable official status
> because there isn't really a good alternative. I think there is a
> need for place=neighborhood, or something like it. Whatever we come
> up with ought to be a solution that handles both the traditional,
> informal notion of a neighborhood, which tends to be relatively
> small and often has nebulous borders, as well as more formal notions
> like the districts of Berlin, or the arrondissements of Paris with
> official status and well defined borders.
>
> -Scott
>
>
> --
> Scott Atwood
>
> Cycle tracks will abound in Utopia. ~H.G. Wells
>
>
> _______________________________________________
> Talk-us mailing list
> Talk-us at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us
----------------------------------------------------------------
michal migurski- mike at stamen.com
415.558.1610
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20081128/9998ee2d/attachment.html>
More information about the Talk-us
mailing list