Change background image in an area created upon an existing area with

Recently I navigated a lot in an area of the wetlands of the Danube Delta (see map=14/45.2438/28.9138) I found many lakes and canals missing, so I created them on the spots where I’ve been and where GPS tracks are available. All well (perhaps I was not very efficient, but I succeeded).

A remaining problem is, that the lakes are defined as an area, specified as lakes, so they become blue. But the underlying “wetlands” background (green with reets, etc, an area created as a set of connected lines) remains partially visible. I found one lake (Purcelu) correct, but that also is defined as a set of lines, not an area.

What should I do to colour the lakes only blue?
I see that many existing lakes nearby are backgrounded identically, perhaps the creator had the same problem but did not bother to be precise?

We seem to be talking about

I would say this is a rendering fault and should be fixed in the renderer, not the map.

However I do note that some of the other lakes have been cut out using inner polygons. I suspect that that is wrong, and that the lakes are as much a part of the wetland reserve as the marshy bits.

You need to find out which is actually true. Either the lakes are part of the reserve, in which case you should fix the renderer, or they are not, in which case you should add them as inner polygons. It is, however, possible, that explicitly making the lakes a layer higher than the surrounding will result in the correct rendering, and, in my view, that would be an acceptable work around. Normally the rule is that smaller areas completely within larger areas implicitly override the larger area, so, for example, you don’t have to cut holes for buildings, or give them an explicit layer.

By the way, you should not think in terms of background images, otherwise you end up mapping for the renderer.

Looked at it better. Relation: Danube Delta Biosphere Reserve (1307091) is defined as an area with outer and inner polygons, actually cutting holes where the major lakes are. Those are blue. All other lakes (defined after the initial import?) are drawn upon the marsh. All lakes are part of the reserve, so your conclusion is correct. I undid my changes and will redo them trying to get better results with rendering with layers.

If you think the data is ok, please report an issue for the renderer in (check first, maybe it’s already reported)

Don’t change what is already good data “trying to get better rendering”. “mapping for the renderer” is not a good practice.

I cannot report an issue. This is my first attempt to add area upon area and I’m quite sure, that I don’t know enough to do it well. Reporting my poor knowledge as an error of the renderer is not what I’m used to do. I was in programming for 45 years and saw this type of shifting responsability too often.
Better: I redid my work. Now do me a favour and go to and zoom in and out, then look if you see an errror in the definition of the lakes and canals that appear and disappear. Different problems, but I guess same origin.
Then report to me, here or as message. Txs.

Appearing and disappearing with resolution is likely to be the result of not allowing enough time for the tiles to be re-rendered (an take hours on the main layer, at times, around a week on some of the others.

Hi Martien,

Try pressing F5 in your browser to refresh the data.

About the proper mapping in the Danube Delta:
Each lake should be an inner polygon to the ‘Danube Delta Biosphere Reserve (1307091)’ relation. If the lake consist of a single line (way), you should add the way to the Danube Delta relation with role ‘inner’. If the perimiter of the lake is build from 2 or more ways, each way should be added to the relation.
I did this for the small lake east from Lake Carasu. However it takes quite some time to render this, because of the size of the Danube Delta relation.
If you look at Lake Purcelu, this lake actually consists of two ways. One to create a hole in the Danune Delta relation (role=inner). And one for the lake itself. Although this is not completely incorrect, it would be better to combine these two way. By doing so, you assure that the hole in the Danube Delta always has the same shape as the lake.

martienvanderburgt wrote “All lakes are part of the reserve,” so it’s clear that the lakes shouldn’t be inner role in the Reserve relation.

In my opinion the relation 1307091 (as an example) is wrongly combining two different things. On the one hand its the boundary of Danube Delta Biosphere Reserve (boundary=protected area etc). The lakes most definitely are part of this Reserve and must not be cut out. On the other hand the relation defines a wetland area (natural=wetland). Lakes are not part of the wetland (well, you could argue they are, but let’s assume they are no for the moment).

The right thing to do would be to split up the relation. Leave one for the wetland and add a new one for the reserve. They could use the same lines in the multipolygon but I seriously doubt that the Reserve’s boundary exactly matches the wetland boundary, not least as wetlands tend to grow and shrink over the years. Actually, there are additional multipolygons and areas that form the reserve, e. g. way 88122556. I guess all these should be combined in the new reserve relation (or way).

edit: typo

@Gertjan Idema
I see what you did. Now I see that lake sometimes yes sometimes not, like before, but in one stage of zooming in and out, the marsh background becomes nicely blue. But one step further again with marsh. Strange things. btw I use Chrome Versiunea 52.0.2743.116 m.

I now see the map correctly on Android, on pc in In IE or Chrome it is not always correct. So after all it’s the renderer.

It is not the renderer, it’s the way images are kept in cache in the browser. It is necessary to refresh each zooming level with F5. Otherwise one sees an old map at certain zooming levels and a new(er) one at other levels. Not simple to understand, but now that I have it, I’ll press F5 often!.