<div class="gmail_quote">On Thu, May 7, 2009 at 9:02 AM, Sebastian Spaeth <span dir="ltr"><<a href="mailto:sebastian@sspaeth.de">sebastian@sspaeth.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="im">Jan Jesse wrote:<br>
><br>
> Hi<br>
><br>
>> In zoom 6 to 11 the tiles of Germany are in a very bad state.<br>
>> Some tiles are completely empty and say to be of "Unknown Type" in<br>
>> openlayers.<br>
><br>
> Just now I tried it again and the problem seems not to be solved?<br>
><br>
> Is anyone working on it?<br>
<br>
</div>There is no procedure to automatically create those caption layers.<br>
Someone needs to feel responsible to either create and upload them,<br>
currently.<br>
<br>
This is how a stock t@h client is able to render the caption layer for a<br>
low zoom tileset:<br>
<br>
./tilesGen.pl xy 34 21 6 caption<br>
<br>
which then needs uploading with<br>
<br>
./tilesGen.pl upload<br>
<br>
Next time that low zoom is stitched it will take the new captions into<br>
account.<br>
<br>
<br>
This is still not optimal. As the supression of labels is based on<br>
proximity of labels we always will get a different selection of captions<br>
for each tileset. The problem gets worse with fewer margin around the<br>
tileset and would be eliminated if we could use the whole world as data<br>
set for our rendering.<br>
I would love to download the caption data from XAPI locally and point<br>
the t@h client to use that local osm file as datasource for all the<br>
caption tiles, but the client is not capable of doing that.<br>
<br>
Downloading the whole world for each tileset is out of question, of course.<br>
</blockquote><div><br>Is it? What captions are you currently downloading? Cities, towns, villages? <br><br>You can grab up to 10 x 10 degrees from XAPI which might be enough to reduce all but the most pathological edge case.<br>
<br>If that isn't enough I could remove that bounding limit for just these tags until we find out what is really needed.<br><br><br>80n<br><br><br></div></div><br>