[Talk-ca] Great Lakes shoreline

[Adam Dunn]

Coastline updates vary, depending on Mapnik vs. Osmarender and what
zoom level you're looking at.

I've done editing to the coastline that represents Great Slave Lake in
the Northwest Territories, and edits made on April 2 still have not
been rendered by Mapnik, but they have been rendered by Osmarender.
Edits made on March 15 are rendered (and have been for a week or two).
So right now it takes around a month for coastline changes to be
rendered by Mapnik.

Adam

On Wed, Apr 20, 2011 at 11:44 AM, James A. Treacy <treacy at debian.org> wrote:
> On Wed, Apr 20, 2011 at 01:58:39PM -0400, Richard Weait wrote:
>> On Wed, Apr 20, 2011 at 12:29 PM, James A. Treacy <treacy at debian.org> wrote:
>> > Hello,
>> > I have been adding canvec data for the last part of the Bruce
>> > Peninsula and noticed that the existing shoreline is quite different
>> > than that given by the canvec data. The source for the existing
>> > coastline is r_coastlines and I have no idea who/what that is.
>> > I don't know which is more accurate but the canvec coastline matches
>> > much better with the land features.
>> >
>> > Should the existing coastline be left alone or should it be switched
>> > over?
>>
>> Dear Jay,
>>
>> That's the eternal question isn't it?  With one source, we just use
>> it.  With multiple sources; it's always about evaluation and
>> comparison. ;-)
>>
>> r_coastlines doesn't ring a bell, but I know that PGS coastlines was
>> fairly easy to improve upon.  Sounds like you are seeing canvec as a
>> better match to the other land features, which may also be from
>> canvec.  I'm not sure you should see that as definitive.  But if the
>> canvec is a better match, and yahoo / Bing don't disagree, ... ?
>
> I just did a comparison with Bing imagery and there is no comparison:
> even given the low resolution of the imagery, the canvec coastline for
> the tip of the Bruce Peninsula is clearly 10x better. Of course, that
> is no indication that the rest of the coastline for Lake Huron is as
> good so would check it as I go along.
>
>> If you decide to proceed, and the water bodies are tagged as
>> coastline, be aware that coastline rendering is sensitive, and does
>> not re-render immediately.
>
> Due to time constraints, I wouldn't start this until May. In fact I
> will have spotty (at best) net connection next week. That is probably
> a good idea as it will give time for feedback from others.
>
> I'd want to start with a test section and see how it goes before
> proceeding. Do you know how long it is between updates of coastlines?
> Also, if it goes well I would just proceed and update a large section
> of the coast. As this would involve a large geographical area, I would
> give updates to the mailing list to minimize the chance of conflicts.
>
> --
> James (Jay) Treacy
> treacy at debian.org
>
> _______________________________________________
> Talk-ca mailing list
> Talk-ca at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-ca
>