[OSM-talk] Group relation proposal
LM_1
intended to reduce tagging duplication and make it easier to map dense
public transport areas by grouping ways that are used by multiple
transport lines (not having to add the same group to multiple route
relations).
The proposal is here:
http://wiki.openstreetmap.org/wiki/Relations/Proposed/Group_Relation
Please discuss or comment, preferably on the wiki discussion page.
Lukáš Matějka (LM_1)
_______________________________________________
talk mailing list
ta...@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk
Richard Mann
Peter Wendorff
I think, the name of the relation is far from optimal, but the basic idea is not the worst, and we already use a similar approach in nested Multipolygon-relations.
But:
1) "type=group" is far too unspecific and misleading, as it's NOT intended to group similar items together (like a category), but to form abstract, unnamed, but common objects to be reused. This is in general the same as forming one outer-area in a multipolygon relation out of several non-closed ways.
2) For the Public transport example I see a major drawback for stops. A common road with several (bus) stops may be shared by different busses, but some the busses may omit different stops where they don't stop. This would require again either to split the common part to several ones (not much better than using the ways directly now), or to use different relations/to use the relation only for some of the routes sharing the same way. But then it does not get that much easier than it is now.
3) Lacking tool support (that's not a good argument, but nevertheless I fear, these relations will for quite a while break most tools and maps using Public transport information, one of the IMHO best showcases of OSM.
regards
Peter
Paul Johnson
On Mar 22, 2012 5:14 AM, "Richard Mann" <richard.man...@gmail.com> wrote:
>
> Relations are not categories. They are for recording geospatial relationships between elements, not for putting things in groups.
I agree, this has the real potential to overcomplicate editing routes in places where which in a large multitude of routes changes relatively regularly over time. TriMet's MAX system, the Portland bus mall, NYC Subway and other big-city systems with even minimal buildup come to mind.
Robert Scott
> I have created a new proposal for group relation (type). It is
> intended to reduce tagging duplication and make it easier to map dense
> public transport areas by grouping ways that are used by multiple
> transport lines (not having to add the same group to multiple route
> relations).
> The proposal is here:
>
> http://wiki.openstreetmap.org/wiki/Relations/Proposed/Group_Relation
IMO this would make life significantly more painful for data consumers.
robert.
Lynn W. Deffenbaugh (Mr)
> On Thursday 22 March 2012, LM_1 wrote:
>> I have created a new proposal for group relation (type). It is
>> intended to reduce tagging duplication and make it easier to map dense
>> public transport areas by grouping ways that are used by multiple
>> transport lines (not having to add the same group to multiple route
>> relations).
>> The proposal is here:
>>
>> http://wiki.openstreetmap.org/wiki/Relations/Proposed/Group_Relation
> IMO this would make life significantly more painful for data consumers.
>
+1 or 1,000 from me! Especially as more developers move to being
data/API consumers rather than tile server retrieve-and-display.
Lynn (D) - KJ4ERJ
Paul Johnson
On Mar 22, 2012 7:32 AM, Petr Morávek [Xificurk] <xifi...@gmail.com> wrote:
>
> Paul Johnson wrote:
> >
> > On Mar 22, 2012 5:14 AM, "Richard Mann"
> > <richard.man...@gmail.com
> So, you don't use it for these routes. Could you explain better how this
> proposal complicates the cases you've mentioned?
Multiplexed routes don't have the same endpoints all the way through. You would have to have multiple group relations to handle it. In the end, this makes it more complex than just mapping each line individually, as is the current method.