On 18/10/2007, <b class="gmail_sendername">Patrick Weber</b> <<a href="mailto:p.weber@ucl.ac.uk">p.weber@ucl.ac.uk</a>> wrote:<div><span class="gmail_quote"></span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Are we just running into a stone wall here of the current server being<br>overwhelmed ? Or can we manage to minimise these inconcistencies by some<br>clever algorithm as proposed?</blockquote><div><br>I'd guess, that short of doing a render of the whole planet and only publishing it each time it 'completes' that you'll struggle to find a way to do this... if you go with the render the 8 surrounding tiles, do you render them all the way down from there? What about the data that borders those 8 tiles, do you also add the surrounding 16, etc? I'd guess that that closest we could get would be by trying to keep all tiles ever rendered, keeping the source data as close to live as possible and only marking tiles dirty that have actually had updates... This is somewhat different to the current model of the mapnik layer... It's reasonably close to what
T@H does though, only missing the 'queue a tile that has changed' step... <br><br>d<br></div><br></div>