Re: [Talk-us] API Behavior

2016-02-26 Thread Simon Poole
Am 26.02.2016 um 18:09 schrieb Richard Welty:
>
> the problem is that detecting the condition is pretty expensive.
> consider long, straight boundaries that pass through your
> bounding box. the search to find these segments of ways
> might have to range out pretty far.
>
>
Pretty far == the whole dataset/the world

Naturally that is simply the extreme case.

 I always recommend to add a node even in completely straight lines
every couple of dozens of meters/yards to be on the safe side and to
keep everything editable.

Forgetting the extreme case, maybe we should have some kind of bounding
box based index for ways and relations (note we don't use PostGIS and it
wouldn't actually help in this case, so this would have to be home grown
in some form), naturally calculating the bounding box would make many
operations that are now very cheap rather expensive.

Simon



signature.asc
Description: OpenPGP digital signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] API Behavior

2016-02-26 Thread Richard Welty
On 2/26/16 12:01 PM, Martijn van Exel wrote:
> It’s an API limitation. If no node is within the bbox the API has no way of 
> knowing there should be something there. I don’t believe there is a different 
> workaround from the one you already suggest yourself, but I would be happy to 
> be wrong (as I too find this an unfortunate limitation).
>
the problem is that detecting the condition is pretty expensive.
consider long, straight boundaries that pass through your
bounding box. the search to find these segments of ways
might have to range out pretty far.

richard

-- 
rwe...@averillpark.net
 Averill Park Networking - GIS & IT Consulting
 OpenStreetMap - PostgreSQL - Linux
 Java - Web Applications - Search




signature.asc
Description: OpenPGP digital signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us


[Talk-us] API Behavior

2016-02-26 Thread Mike Thompson
When using the standard API to download an area to edit (through JOSM) it
seems that if a way runs through the bounding box you provide the API, but
doesn't have any nodes in that bounding box, it will not be returned.  This
can result in collecting the same data a second time because the first
instance is not returned.

1) Is this a JOSM issue or an API issue?
2) Is there any work around other than to download using a larger bounding
box and hope that the way in question references at least one node in that
bounding box?

Mike
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us