NOTICE: is no longer in use from 1st March 2024. Please use the OpenStreetMap Community Forum

In the digital world, when viewing/rendering objects, the smallest unit is a pixel. Consequently, the border lines of areas defined by MPs are one pixel thick lines. So, the decision whether this one pixel (inner) border line belongs to the area or not may have serious rendering and other functional consequences.

asked 17 Nov '18, 10:23

sanser's gravatar image

accept rate: 5%

With vector graphics (or indeed map data) the pixel is not necessarily the smallest unit. While the data itself will have a limited precision, it doesn't necessarily have a resolution in the sense used in raster graphics.

While in many cases it is traditional to draw an outline separately on top of the border between e.g. inside and outside, this is not necessary for all rendering systems. The choice may be made to treat the line as having infinitesimal width and define the inside to be one colour and the outside another. If this then needs to rasterised for display then a variety of techniques may be used to determine the final rendered colour. This may be done very simply by, for example, determining whether the centre of each pixel is on one side of the line or the other and then colouring the entire pixel according to that determination, or it may be done by more sophisticated means whereby a determination is made as to how much of the area covered by the pixel is inside or out and choosing a colour for that pixel as being somewhere in between fully inside and fully outside.

When it comes to OpenStreetMap data, remember that there is no unified size of a pixel, this will depend on the renderer, zoom level etc. AFAIK OSM has no nominal width to outline features even though some objects recorded as centrelines can have explicitly defined widths. To me, trying to make allowances for the dividing line itself potentially having a displayed width strays into "tagging tracing for the renderer".

permanent link

answered 17 Nov '18, 12:01

InsertUser's gravatar image

accept rate: 19%

Of course. The first lesson in vector/raster based imaging. However the question is much more serious than that. From the many "functional consequences" let us take this illustration. In OSM neighbouring holes in MPs are absolutely legal. If the "inner" border belongs to the area the common border section of the neighbouring holes is part of the area and the contrary. This dilemma has serious topological/connectivity consequences no matter whether you draw or not the area and which resolution (penn thickness) you assume.

(30 Nov '18, 09:00) sanser

To my mind, the edges of areas have zero with, if as in your example two inner polygons have share a line then the actual inner is the union of the two areas and there is no area between the two regions. To make your example more concrete: if a forest surrounds a grassy clearing and lake and the grassy area borders the lake, the grassy area and lake may both be listed as inner to the forest, but it would be wrong to assume that there is a thin line of forest separating the two entities.

(30 Nov '18, 09:27) InsertUser

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here



Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text]( "title")
  • image?![alt text](/path/img.jpg "title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Question tags:


question asked: 17 Nov '18, 10:23

question was seen: 1,678 times

last updated: 30 Nov '18, 09:27

NOTICE: is no longer in use from 1st March 2024. Please use the OpenStreetMap Community Forum