No one has undone any mapping work here, mapwitch has just split a non-maintainable giant multipolygon stretching over thousands of kilometers, without doubts the largest riverbank polygon in the database, into smaller, easy to handle parts. There is no loss in information and mapping work involved in that. This is completely in line with established mapping practice world wide and a very positive contribution making future edits in the area much easier.

Riverbank polygons should always be split into relatively small, easy to handle parts. Large riverbank polygons are difficult to edit and to properly verify for validity for the mapper and more difficult to handle for the data user, they break much more frequently than small polygons and with much more severe consequences.

Thanks. That's a very good deed.

Hello ronic,

thanks for your contributions to Antarctic mapping.

You seem to have problems choosing the right tags for what you map. place=neighbourhood indicates a populated area in an urban environment, it is not a fitting choice for tagging physical geography features. natural=saddle (osm.org/node/2335535216) is not compatible with something located directly at the coast. If you are unsure about choice of tags for certain features (which is not uncommon in an exotic environment like here) just ask on one of the OSM communication channels like mailing lists or forum in your preferred language or at least add a description tag explaining what kind of feature you want to map here.

A Bluff (like in osm.org/node/5537488267) is an exposed rock area or cliff, you can tag the name on a natural=bare_rock area (which already exists here - osm.org/relation/2968576) or draw a line along the edge of the cliff and tag it natural=cliff.

Dass sich künstlich und natürlich nicht unterscheiden lassen ist ein gelegentlich in Analogie vorgebrachtes Argument, das greift aber bei Fließgewässern im Grunde nicht. Natürlicher Wasserlauf bedeutet nicht naturbelassen, sondern dass seine Existenz natürlich bedingt ist.

Im Flachland ist das nicht immer ganz einfach zu erkennen, im Gebirge aber meist schon, insbesondere wenn man das Relief mit betrachtet - welches ja durch die natürlichen Wasserläufe entscheidend geformt wird.

Und die Benennung eines Baches als Graben ist wie gesagt kein Indiz für eine künstliche Anlage. In der Umgebung der Wutachschlucht rühren solche Namen vermutlich einfach daher, dass sich die Bäche in den Untergrund eingraben.

Hallo BeKri,

bitte achte darauf, kleine natürliche Wasserläufe immer als waterway=stream zu taggen, waterway=ditch ist ausschließlich für künstlich angelegte Gräben. Das ist unabhängig von der Benennung - dass etwas 'Graben' heißt bedeutet nicht, dass unbedingt es sich um einen waterway=ditch handelt.

Grüße,

Christoph

Hello Carolina,

could you explain what you are trying to do here with

osm.org/way/520138444
osm.org/way/44036775
osm.org/relation/7526598

This clashes with the purposes of natural=water and natural=coastline.

Hello visaman-import,

could you point me to where this import and its tagging have been discussed and are documented?

The data of this and other changesets from you in the same area uses undocumented and partly redundant tags, URLs that redirect to an empty page and the data in parts contradicts existing data in OSM.

Based on

osm.wiki/One_feature,_one_OSM_element

i deleted nodes for reefs that are also mapped as areas - after transferring tags to the area.

natural=tundra does not say anything about the vegetation except for that it is without trees. I would advise against it. Grassland is completely fine here for areas with vegetation.

Thanks.

If you can reliably identify differences in the vegetation from available images it does not hurt to differentiate them with supplemental tags (like grassland=*).

Hello dikkeknodel,

nice mapping work on South Georgia - but you should probably be aware that there is no scrub vegetation around there - to my knowledge there are hardly any woody plants growing on South Georgia. What you identify as scrubland on images is probably mostly tussock grass formations.

See

http://nora.nerc.ac.uk/id/eprint/509458/1/The%20vascular%20flora%20of%20South%20Georgia%20-%20BAS%20Scientific%20Report%2045.pdf

for background info on the vegetation of South Georgia.

This is just my assessment of the situation - i won't insist on changing anything here. But i am pretty sure the position is wrong, not wrong by a few hundred meters but wrong by likely many kilometers - either much closer to the shore or much further to the east (that is the problem with secondary sources - you have this kind of error in Wikipedia all the time).

Regarding future improvements - it is unlikely that when suitable images are available for the area such a volatile feature still exists.

Hello Dave,

i would advise against trying to map features where you have nothing even close to verifiable position data for.

The location of node 5125769091 is a highly unlikely position for an islet or even a gravel bar based on the sea ice movement in the area - ice would not be able to move across such a feature freely - but it does in the area.

There are lots of well verifiable features missing in the north of Greenland in OSM - lets concentrate on those and not on data of questionable reliability like this.

velmyshanovnyi,

you need to verify your edits more diligently otherwise this can only be seen as a mechanical edit/import. For example at:

https://mc.bbbike.org/mc/?lon=-153.046905&lat=70.038537&zoom=12&num=2&mt0=bing-satellite&mt1=mapnik

the islands in the lake are actually clouds and residual ice. If you familiarize yourself with the region and use other readily available image sources for verification you can see that. You don't do anyone a favor by mechanically tracing thousands of lakes this way without checking for such errors.

And you should definitely not use the 15 year old low resolution images in Bing at z12 and below as a basis. Both Bing and DigitalGlobe have higher resolution and newer images in most of the areas you are working on. Make use of that. And it might be good to ask other experienced scanaerial users like Vort and Jake Strine for help with best practice in using such techniques.

The tags added indicate these edits might be based on a tool like scanaerial which would mean this might not actually be a mechanical edit.

However this should still be discussed before widespread use because (a) the quality of the geometries is poor indicating a rather crude algorithm behind it and (b) this seems to be based on poor and outdated lowzoom Bing images which differ a lot from the current reality - see for example here: https://mc.bbbike.org/mc/?lon=-152.654172&lat=70.015963&zoom=13&num=2&mt0=bing-satellite&mt1=mapnik

Bonjour abdeldjalil,

i am sorry for missing the existing mountain region polygon.

However the tagging is still quite wrong since in the area covered exposed bedrock is not the dominating surface type. Most of this area is covered by loose material - mostly either scree or alluvial deposits - like here: https://mc.bbbike.org/mc/?lon=8.069623&lat=25.976448&zoom=14&num=3&mt0=bing-satellite&mt1=mapnik&mt2=google-satellite

Exposed bedrock is mostly limited to steep parts where relief prevents accumulation of loose material.

Having such a big polygon covering a very heterogeneous area makes it difficult for mappers to correctly map details more specifically on a local level.

Merci,

Christoph

Hello Andre68,

this edit is factually incorrect and creates a several thousand square kilometer error in the coastline. Could you please restore correct tagging placing the coastline where the coastline actually is an tagging the bay with the established tag for bays (natural=bay - osm.wiki/Tag:natural%3Dbay)?

Hello tasauf1980,

i am not sure what you are trying to do here but it is not working. You have added five new revisions to osm.org/way/100827142 and the end result is both geometrically (due to self intersections) and semantically incorrect.

If a large and complex multipolygon relation is difficult to handle for you you can split the water area into smaller simple polygons - as it has been done before you created the multipolygon in osm.org/changeset/51275200

And please use meaningful changeset comments so others can understand what you are doing - see osm.wiki/Good_changeset_comments

Hello palimpadum,
not sure if you are aware that you are refining glacier mapping here based on images which are probably older than the ones these were originally mapped from three years ago and which are partly off by more than 100m.