[OSM-dev] Very long ways have been split

Frederik Ramm frederik at remote.org
Tue May 15 01:16:51 BST 2007


> I think this earlier thread gives a few more details about the process
> used
> http://lists.openstreetmap.org/pipermail/talk/2007-February/010879.html
> It seems no one else bought into the idea of moving the coastline data
> out of the OSM DB.

I remember having discussed this with Martijn on the list. There were 
(vague) plans to create costal and ocean tiles that could be used either 
as a "background" layer for normally rendered tiles or merged with these 
in a bitmap operation, so that OSM could be completely ignorant of where 
the coastline is.

Martijn was in favour of such a solution, saying that it is very 
inefficient to keep all that static data in the database (and render it 
every time!).

My counter-argument was that (1) I want to see the coastline while I am 
editing, and (2) I want to have the power to move it, right there and 
right then, in my editor, if it seems wrong. Having the data in some 
sort of external repository where mappers cannot change it (or only with 
special tools) seemed to me so un-OSM-ish.

When I implemented the coastline stuff for tiles at home - after discussing 
it with Deelkar and others in Essen - I shamelessly made use of the OSM 
principle "he who does the job gets to decide". There were complaints 
initially (hello Lars), but since I did that, anyone with JOSM can now 
do coastline work, and see the results quickly. People seem to have 
bought into this idea. In the end, everyone will benefit from the 
high-quality coastline data that is being created this way.

Even if we decide to move away from having the data in our database (and 
I could well imagine coastline data residing on a different server, with 
a matching JOSM plugin to work on it, and rendering special tiles from 
there), we'll at least have high quality data then.


Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00.09' E008°23.33'

More information about the dev mailing list