Re: [aprsisce] Tiles and Oceans/Seas
- I'd prefer to use a fixed blue-ish color for all the fake tiles. That way there is no question that it's not real, in case at some zoom you happen to catch a spec of something in the single tile you happen to grab. Also, using a different color than the default OSM "Ocean" tile will make it clear where the boundary of real and fake is, again, just in case you were looking for that spec of something in the real tile, just off the edge of a fake one.
Pretty isn't pretty if it gets in the way with being able to trust what you see.
Greg KO6THOn Wed, Nov 30, 2011 at 4:25 PM, Steve Daniels <steve@...> wrote:
Yes just grab a single tile from within the area, if you have not already, or not, as a not matching colour may not look pretty, but would show a not mapped area.
I can start on boxes for ocean/sea areas that have no land in them.
How far do we go with it? Obviously things like the major Oceans, and large areas of Sea, but maybe a series of boxes up the English Channel is going a bit far?
The idea obviously is to not waste bandwidth downloading essentially useless tiles.
I don’t know enough about vector data to have an opinion, I know it would be a lot smaller, and you could probably have the whole world in a releativly small file.
I got side tracked, I have had the privilege of working with some artists, before they stopped doing things like being machinists, and got recgonised for the art.
My boss in the early 90’s did art using vector graphics, just looked up his art.
Torbay Freecycle Moderator
From: firstname.lastname@example.org [mailto:email@example.com] On Behalf Of Lynn W Deffenbaugh (Mr)
Sent: 30 November 2011 23:16
Subject: Re: [aprsisce] Tiles and Oceans/Seas
And the really nice thing about "all blue" rectangles is that they could serve up the "blue" tile at any zoom level.
However, who's to say what color "blue" the "ocean" is supposed to be? Doesn't that depend on the tile set that's in use? Hmm, maybe just fetch a single tile from within a rectangle the first time the rectangle is hit for a given tile set and latch onto that color for that tile set's "blue"... And clear the "blue" color whenever the tile set is edited because we won't know if the user actually changed tile sources within a single set. Oh, and only latch onto the "blue" if the fetched tile is a single color...
Oh, and to use a substitute "blue" tile, the ENTIRE desired tile must be within a single blue rectangle or the tile will be fetched from the server. This is to prevent further-out zoom levels that have land and ocean from accidentally using a pure blue tile.
So, any volunteers to develop a/latN/lonW/latS/lonE rectangles (see APRS-IS filter specs) of pure blue areas? Better yet, is there anyone conversant enough with OSM vector data that can either a) generate such rectangles from OSM data or b) validate that there's no ways or nodes within rectangles that someone else has come up with?
Lynn (D) - KJ4ERJ - Author of APRSISCE for Windows Mobile and Win32
On 11/30/2011 6:04 PM, Steve Daniels wrote:
Due to a bit of confusion Lynn and I were discussing OSM and tiles tonight, (I asked about tle updating and Lynn read that as tile) on that subject I suggested auto updating of tle’s
Lynn is welcome to repost relevant parts of the chat.
I started of thinking about torrents, that is having an inbuilt torrent system for sharing tiles to take the load off the OSM servers, with new installs maybe getting a preliminary file of their country.
But that does require people to allow connections through their routers.
My main thought though was a lot of us download lots of useless tiles, for example tracking a satellite across the pacific, or a boat etc.
So I got around eventually to thinking of perhaps creating co ordinate boxes, of areas of ocean/sea with no land in them, that way APRSIS could serve up blue tiles without downloading them, for those areas
Torbay Freecycle Moderator