On 2010-08-22, Norbert Hoffmann wrote:
> Marcel Lohmann wrote:
> >can someone explain me, why is it necessary to render empty tiles?

Think ferry lines. Think captions that overlap from a neigboring tile
into the empty tile. Just because there is no data in our tile it does
not necessarily mean there will be nothing rendered in it.

And if it really is empty, Norbert is right, the result will be used to
wipe out any data that might have been there before.

> - the tile has some new data added to the database, but the key-value has
> no rendering rule for it,

Right, that would also be a possibility.

There is only an automatic request for rendering tiles if data has
changed on it, so t...@h usually does not issue render requests for empty
tiles. The old-tiles checker that petschge runs will probably rerequest
empty tiles from time to time, but I don't know how often that is.

That having said, there are certainly shortcuts that we could take on
the client side when the client recognizes that a tile is really
empty. If there is no data in a tileset, it could e.g. immediately
upload an empty tileset tile, rather than creating one from scratch
every time, needlessly optimizing png files and stuff.

Sebastian

_______________________________________________
Tilesathome mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/tilesathome

Reply via email to