[Imports] Tasmanian Parks

[Jason Remillard]

Hi Ian,

Obviously It is not wrong to include the source and attribute in the
data. It is used all over the place. However, consider when the data
is modified after your import, what are people supposed to do with
those tags? The data provider might be offended if the data is sourced
from them, but was actually hand edited against some other source.

You should look around the area and see how things are tagged, I think
it might vary region to region. The only issue I see with what you
posted is the protection_title values are not present in the taginfo,
they are new. If might be ok, just wanted to point that out.

Also another thing to keep an eye out for is how intense the relations
are. If the imported data is a big hairball of relations it can be
very difficult to edit it afterwards. This seems to be an issue for
land use imports.

It would be ideal if the completed OSM file could be posted before you upload.

Thanks
Jason.


On Tue, Jul 2, 2013 at 4:18 PM,  <porjo38 at yahoo.com.au> wrote:
>
>
> Hi Jason,
>
> Putting attribution & source tags only on the changeset seems wrong to me. When a 3rd party takes the data from OSM, the attribution at least should be attached, no?
>
> I'm using ogr2osm to generate XML. Below are a couple of samples - first is RES_CLASS="Conservation Area" and second is RES_CLASS="Forest Reserve"
>
> <relationid="-574116"visible="true"><memberref="-574117"role="outer"type="way"/><memberref="-574667"role="inner"type="way"/><memberref="-574693"role="inner"type="way"/><memberref="-574713"role="inner"type="way"/><tagk="type"v="multipolygon"/><tagk="attribution"v="DPIPWE,  http://www.dpiw.tas.gov.au/tasreserveestate"/><tagk="name"v="White Beach Conservation Area"/><tagk="leisure"v="nature_reserve"/><tagk="source"v="Tasmanian Reserve Estate Spatial Layer 2012"/><tagk="operator"v="DPIPWE"/><tagk="boundary"v="protected_area"/><tagk="start_date"v="2011-10-12"/><tagk="protection_title"v="Conservation Area"/></relation>
> <relationid="-50546"visible="true"><memberref="-50547"role="outer"type="way"/><memberref="-50985"role="inner"type="way"/><tagk="type"v="multipolygon"/><tagk="attribution"v="DPIPWE,  http://www.dpiw.tas.gov.au/tasreserveestate"/><tagk="name"v="Derby Forest Reserve"/><tagk="source"v="Tasmanian Reserve Estate Spatial Layer 2012"/><tagk="operator"v="DPIPWE"/><tagk="boundary"v="landuse"/><tagk="landuse"v="forest"/><tagk="start_date"v="1998-12-18"/><tagk="protection_title"v="Forest Reserve"/></relation>
>
> Look OK? Is boundary type 'landuse' correct for a forest reserve?
>
> Regarding tagging of entities, I'm thinking that only relations need to be tagged and ways that are not part of relations. Is that correct?
>
> My planned workflow is:
>
> For each RES_CLASS:
>
> - use ogr2sm to produce XML
> - load into JOSM to simplify ways and tidy up
> - bulk upload from JOSM using account 'porjo_tasparks'
>
>
> Ian.
>
> ________________________________
> From: Jason Remillard <remillard.jason at gmail.com>
> To: IanB <porjo38 at yahoo.com.au>
> Cc: Imports OpenStreetMap.org <imports at openstreetmap.org>
> Sent: Sunday, 23 June 2013 11:29 PM
> Subject: Re: [Imports] Tasmanian Parks
>
>
>
> Hi Ian,
>
>
>
> On Sun, Jun 23, 2013 at 4:03 PM, IanB <porjo38 at yahoo.com.au> wrote:
>
> Thanks for the advice.
>>
>>
>>> - Put your static tags on the change set.
>>
>>
> Do you mean these should go *only* on the change set?
>
> Yes.
>
>
>>
>>> - I don't see any tags that actually identify the features being imported?
>>
>>
> I'm not sure what you mean. The 'name' tag will identify them...but I guess you meant something else?
>
>
> The "RES_CLASS" field  should probably map to different OSM tags. For example, a national park does not get the same tags as a historic site. If you are just importing national parks, and not the other RES_CLASS's, then you are missing the "boundary=national_park" or "boundary=protected_area", which is the primary tag over the protection_title tag.
>
>
>
>>
>>> - You might need an area tag on some of these features?
>>
>>
> Providing that all features are in multi-polygon relations, no area tag should be needed?
>>
>>Ian.
>
> Yes, you might be right. In general that