Steve Ratcliffe wrote:

...

Hi

...

points style:

shop=car_repair [0x2c14 continue] shop=car_repair [0x2c09]

and a polygon style:

shop=car_repair [0x12]

Processing this on a very simple OSM file containing one polygon tagged shop=car_repair and one point tagged shop=car_repair, using the --add-pois-to-areas switch, I see: One polygon of type 0x12, with a 0x2c09 POI in the centre Two POIs consisting of one of type 0x2c09 overlain on top of another of type 0x2c14

Yes that does appear to be the case.

It might be difficult to change. We don't know if something is a polygon until after the style rules have been run and at that stage it is too late for the normal processing to occur for the generated point. At least that the way it is written - it is possible that it could be changed, if anyone has any ideas please post.

Regards,

..Steve

I'm sure I saw an email from WanMil recently that said there already is code in mkgmap that can test to see if an object is a closed polygon.

Yes, it is possible to check if an object is a closed which does not necessarily mean that it is a polygon (example: roundabout). But we could add some methods (maybe they already exist) that can check if an object is tagged with polygon tags.

If, so, then this code could be re-used and --add-pois-to-areas could be run before the style file processing. Whilst I'm helpfully palming off work to others more skilful than myself, could whoever does this also add some logic to make sure the resultant POI is always placed *inside* the polygon? :)

Do you have an algorithm that performs within an acceptable time?

-- Charlie _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

On 16.10.2010 13:36, Walter Schlögl wrote:

Hi,

the feature add-pois-to-areas seems to have a disadvantage, not working together with the continue command.

Altough I did not fully understand the reason, (maybe it is related to the ordering of processing polygons and pois, maybe it's because of difficulty to identify closed polygons) I will nevertheless try to give a proposal for a future solution.

At the moment all add-pois-to-areas have to be identified by processing the polygons style file, and all items having no style here will not get a POI mapped.

My idea is a special style file that is processed before polygons and points. The content of this "substitute" style file shall be used to 1.) define polygons that shall be added as poi (in the center), 2.) define lines that shall be added as poi (in the center or at both ends), 3.) include all name changes on only one location.

I'ld fully support this. The current behaviour is simply not very practical. Other possibility would be a a key to specify the behaviour inside the polygons file (like a) only polygon b) only POI c) POI+Polygon)...

Torsten Leistikow wrote:

As a user I would like to have an action command in the polygon (and also in the lines) style rules for generating a POI for a specific area (and also for a specific line). Actually I would like to have multiple commands, e.g. add-centre-node (new node in the geometric centre of the polygon) add-start-node (first node of the polygon) add-end-node (last node of the polygon) add-middle-node (middle node of the polygon) add-inner-node (all nodes of the polygon beside the first and the last one) add-all-node (all nodes of the polygon)

I could think of multiple uses for such a feature.

I do not know, whether this would be possible, since it would require a processing of the polygon and line rules before the processing of the point rules.

+1 I think, you'll need a preprocessing, searching for all closed ways an create new Nodes at the required position. After this mkgmap can run in normal order aighes -- View this message in context: http://gis.638310.n2.nabble.com/Re-continue-command-in-combination-with-add-... Sent from the Mkgmap Development mailing list archive at Nabble.com.