I totally agree with you on that. I'm one of Glenn's beta testers and
I have noticed that depending on the town, the data given by GRB/Crab
can be good or very bad/outdated.
Also the data is rather limited in the types of buildings that it
recognizes, which is a problem for me.
Feel free to take a look at [1] to see stuff that is obviously wrong
in GRB, even from aerial imagery. I gave up on adding more screenshots
recently, but it should give you an idea of the data quality you can
expect :-)


m.


[1] https://xian.smugmug.com/OSM/Screenshots/AGIV-Problems/

On Fri, Jul 1, 2016 at 4:08 PM, Gerard Vanderveken <g...@ghia.eu> wrote:
> Congrats!!!
> ( I didn't  connect the OSM username with you.)
>
> To me, as there are not much real unmapped places in Belgium, you should
> always start from the map and during survey check what is missing and
> otherwise, if what is present on the map is actually still present and
> correct.
> Else you could take a lot of pictures and notes of things already done.
> Seems a waste of time and effort to me.
>
> As said below, for me the current mapping of sidewalks is fine!
>
> For the buildings and house numbering, I see a lot of the same problems with
> these automatic imports as in regard with the 'import' of satelite photo's.
> Every day, buildings, new streets, ... get constructed or demolished.
> Satelite photo's and databases are frequently not on par with the real
> situation.
> eg. In Huldenberg you have Zagerijstraat, Priesterberg, 20 houses down at
> the St-Jansbergstwg, etc.
> IMO, every import at the desk, should be followed or preceeded by a survey
> at place, to verify current status.
> I consider these tools more as a drawing aid and see the real force of the
> open street map in its continual survey by the many and as such following
> much closer the reality, then the other tools.
>
> Regards,
> Gerard
>
>
> Marc Gemis wrote:
>
> I know I mapped the place in detail, but I hoped people would mention
> what they would map without looking at the map :-)
>
> I stopped mapping house numbers ( I have collected more than 20.000 by
> hand according to Pascal Neis :-)  ), as they are imported via CRAB at
> the moment and maybe later this year via a GRB import (with the
> buildings).
> That's why I turned to other stuff.
>
> I've heard pro's and con's for separate sidewalks & cycleways. I don't
> separate them as long as there is no physical barrier (hedge, guard
> rail, ditch,...)
> I'll try to map sidewalk=... or cycleway=... on the road though.
>
> I'll have a closer look at the TIA sign
>
> m.
>
> On Fri, Jul 1, 2016 at 2:53 PM, Gerard Vanderveken <g...@ghia.eu> wrote:
>
>
> Hi,
>
> Checking on the map, learns me that most items are already thoroughly mapped
> (street cabinets, bins, advertising, lanes, crossings,...)
> Things lacking, but not derivable from the photo are the division of
> buildings and the corresponding addresses of the buildings.
> IMO addresses are of more importance then the little street furniture and
> should get more priority in mapping.
> Not meaning, that I don't appreciate these mappings, every added object
> counts!
>
> When you have lot of time, you could also map the separated footwalk and
> bicycle paths, but this is advanced stuff, and I don't recommend it, as it
> is not so easy to make a correct separation and have all options for
> connections and resulting in no foot etc on the main highway itself, so that
> routing programs for bike and foot will always select the correct paths. If
> not done well, it makes more clutter then making things real or clear.
> It seems these paths are already good and sufficient tagged in the map.
>
> So, in short, at first sight, not much extra to map from the photo.
> (I assume the sign to TIA is simply a direction and not part of a (tourist)
> route, which would imply a missing relation for some bicycle route/network)
>
> Regards,
> Gerard.
>
> Marc Gemis wrote:
>
>
>
> I wonder whether we could learn to see/map more by looking at each
> other's survey pictures. Would it allow novices to "see more" ? Let's
> try out
> I know the mailing list does not allow to include pictures, so you
> have to do it with a link
>
> What could/would you map when you look at [1] ? What would you
> investigate further ?
>
> Feel free to answer in French/German/English or Dutch
>
> p.s. Feel free to look at the other pictures at the site [2] and map
> anything you want :-) Under the picture you'll find a (i) which shows
> a map when you click it
>
>
> ---- Nederlands ----
>
> Ik vraag me af of we van elkaar kunnen leren als we kijken naar de
> foto's die we maken tijdens een survey ? Zou het helpen om
> (beginnende) mappers meer te leren zien ? Laat ons het eens
> uitproberen. Jammer genoeg laat de mailing list niet toe om foto's toe
> te voegen, dus moet je het met een link stellen
>
> Wat zou je kunnen mappen, of wat map je gewoonlijk als je [1] bekijkt
> ? Wat zou je van nader bij gaan bekijken ?
>
>
> p.s. Het staat je vrij om naar de andere foto's op de site [2] te
> kijken en om het even wat te mappen dat je erop ziet :-) Onder de foto
> staat een (i), als je daarop klikt, zie je een kaartje.
>
>
>
>
> [1]
> https://photos.smugmug.com/OSM/OSM-2016/2016-06-09-Wilrijk/i-b8Xdd5g/0/O/DSC_3228.jpg
> [2] gallery:
> https://xian.smugmug.com/OSM/OSM-2016/2016-06-09-Wilrijk/i-b8Xdd5g
>
> _______________________________________________
> Talk-be mailing list
> Talk-be@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-be
>
>
>
> _______________________________________________
> Talk-be mailing list
> Talk-be@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-be
>
>
> _______________________________________________
> Talk-be mailing list
> Talk-be@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-be
>
>
>
> _______________________________________________
> Talk-be mailing list
> Talk-be@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-be
>

_______________________________________________
Talk-be mailing list
Talk-be@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-be

Reply via email to