Re: [Talk-ca] Discussion: zones boisées
Bonjour Dega, En effet toutes les clairières devraient être taguées en consequence par un polygone. Un tag approprié pour cela pourrait être meadow ou scrub. Quand à la présence ou non d'un sentier... difficile à dire par la vue satelite: il faut vraiment un reperage terrain par des contributeurs locaux. D'où la complémentarité entre une communauté locale vivante et des données Canvec qui ne fournissent qu'un matériel brut Bruno Le 2014-11-20 10:30, Ga Delap gade...@gmail.com a écrit : Bonjour Bruno Je suis d'accord avec la logique de Frank : redécouper les polygones de CanVec selon les lignes de frontiéres naturelles (rivières, lacs...) et artificielles (routes, lignes électriques, lignes coupe-feu...etc..). J'ai regardé tes exemples. C'est un compromis intéressant. La zone fôret est séparée en parcelles délimitées par toute entité linéaire (route, ligne électrique, voie ferrée, etc) ou surface (lac, bâtiment, clairière, etc). Je n'ai vu aucun trou. Mais il ne faut pas oublier de définir l'entité qui se trouve entre 2 parcelles car l'absence de forêt n'est pas, en soi, une entité. Par exemple, il y a surement un objet physique qui sépare les chemins 302348523 et 302349825. Je ne connais pas le territoire mais ce semble être un chemin forestier qu'il faudrait définir. Et, à la limite, ce chemin linéaire devrait être défini à l'intérieur d'une surface 2D de type clairière. Ce qui m'améne à poser de nouveau une question de l'article qui a initié cette discussion: Est-ce correct d'utiliser l'absence de définition (undefined) pour représenter les clairières? Est-ce que toutes les zones blanches de la carte OSM sont des clairières? Bonne journée dega ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
Re: [Talk-ca] Discussion: zones boisées
Hello Dega / all, With regard to a path that cuts through a forest, the initial thought might be to map the path as dirt or something similar. However, this would not be efficient nor required. A line that denotes a path is, in itself, technically a type of land tag as the any such pathway must meet some basic criteria to be a path (deliberate connection, open land area, clearly defined, etc). For example, a regular residential road that cuts through a residential area. Do we designate the entire area, including the land that the roads are on as residential or do we segregate the land that the road occupies as something separate from the residential land use? The first one allows for large, uninterrupted polygons denoting a residential area while the second makes for dozens of small polygons separated by roadways. It all depends on what you think the line that makes up the road itself on the map represents. If it represents a type of land use tag, then the first case makes sense as the land is residential in general, except for any area marked with a route as that route would mark the land as used for a road within the residential area. If roads do not tag the ground beneath it, then we need to specifically set what the ground underneath is to be tagged as, requiring multiple smaller polygons to map. Adam 2014-11-21 9:16 GMT-03:30 Bruno Remy bremy.qc...@gmail.com: Bonjour Dega, En effet toutes les clairières devraient être taguées en consequence par un polygone. Un tag approprié pour cela pourrait être meadow ou scrub. Quand à la présence ou non d'un sentier... difficile à dire par la vue satelite: il faut vraiment un reperage terrain par des contributeurs locaux. D'où la complémentarité entre une communauté locale vivante et des données Canvec qui ne fournissent qu'un matériel brut Bruno Le 2014-11-20 10:30, Ga Delap gade...@gmail.com a écrit : Bonjour Bruno Je suis d'accord avec la logique de Frank : redécouper les polygones de CanVec selon les lignes de frontiéres naturelles (rivières, lacs...) et artificielles (routes, lignes électriques, lignes coupe-feu...etc..). J'ai regardé tes exemples. C'est un compromis intéressant. La zone fôret est séparée en parcelles délimitées par toute entité linéaire (route, ligne électrique, voie ferrée, etc) ou surface (lac, bâtiment, clairière, etc). Je n'ai vu aucun trou. Mais il ne faut pas oublier de définir l'entité qui se trouve entre 2 parcelles car l'absence de forêt n'est pas, en soi, une entité. Par exemple, il y a surement un objet physique qui sépare les chemins 302348523 et 302349825. Je ne connais pas le territoire mais ce semble être un chemin forestier qu'il faudrait définir. Et, à la limite, ce chemin linéaire devrait être défini à l'intérieur d'une surface 2D de type clairière. Ce qui m'améne à poser de nouveau une question de l'article qui a initié cette discussion: Est-ce correct d'utiliser l'absence de définition (undefined) pour représenter les clairières? Est-ce que toutes les zones blanches de la carte OSM sont des clairières? Bonne journée dega ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
Re: [Talk-ca] Discussion: zones boisées
On Fri Nov 21 2014 at 8:05:08 AM Adam Martin s.adam.mar...@gmail.com wrote: It all depends on what you think the line that makes up the road itself on the map represents. If it represents a type of land use tag, then the first case makes sense as the land is residential in general, except for any area marked with a route as that route would mark the land as used for a road within the residential area. If roads do not tag the ground beneath it, then we need to specifically set what the ground underneath is to be tagged as, requiring multiple smaller polygons to map. I've seen this question pop up again and again, especially in countries with a lot of small parcels of different land usage. For example, In Germany it's very common to have several different land uses along one road. So do we attach them via node to the way of the road or do they stop x meters from the centerline of the road? What if you have trees overhanging a road? How wide and developed has a trail have to be for it to stop being part of a forest and become its own land use? How do you deal with the editing nightmare of many different land uses glued to a road? From what I've seen, there are good reasons for both perspectives and it is highly unlikely that there will be consensus (or even a clear majority) for either view in the community. Harald. ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
[Talk-ca] Vancouver Meetup
We are having our second meetup on Tuesday, November 25 in Vancouver. The meeting will be held at the Community Living Society, 713 Columbia Street, New Westminster, BC. Sign up on http://www.meetup.com/OpenStreetMap-Vancouver/events/218377932/ to attend. It will be a great time. Don't miss it. Clifford -- @osm_seattle osm_seattle.snowandsnow.us OpenStreetMap: Maps with a human touch ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
Re: [Talk-ca] Discussion: zones boisées
Hello, Yes indeed some European countries are going further into smaller details. And it's legit. However, in the Canadian context, it seems that adjust fix massive CanVec data imports has higher priority than touching up every single landuse parcel It's not only about woodlands water, but also some conversion errors like Canvec schools converted into jails in OSM. So every single little village in QC has his own little jail ! Funny? ;) (Some of them have been fixed...) And double-space in street names is also a pain when using Nominatim Zero un-taged pixel world map is high hope :-) Greatings ! Le 2014-11-21 09:42, Harald Kliems kli...@gmail.com a écrit : On Fri Nov 21 2014 at 8:05:08 AM Adam Martin s.adam.mar...@gmail.com wrote: It all depends on what you think the line that makes up the road itself on the map represents. If it represents a type of land use tag, then the first case makes sense as the land is residential in general, except for any area marked with a route as that route would mark the land as used for a road within the residential area. If roads do not tag the ground beneath it, then we need to specifically set what the ground underneath is to be tagged as, requiring multiple smaller polygons to map. I've seen this question pop up again and again, especially in countries with a lot of small parcels of different land usage. For example, In Germany it's very common to have several different land uses along one road. So do we attach them via node to the way of the road or do they stop x meters from the centerline of the road? What if you have trees overhanging a road? How wide and developed has a trail have to be for it to stop being part of a forest and become its own land use? How do you deal with the editing nightmare of many different land uses glued to a road? From what I've seen, there are good reasons for both perspectives and it is highly unlikely that there will be consensus (or even a clear majority) for either view in the community. Harald. ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca