[OSM-talk] Mapnik has busstops

[Steve Chilton]

Let's not forget that this is a cooperative/collaborative project. The Osmarender and Mapnik outputs should not be seen as in competition but as two parallel parts of the project that have their own strengths and weaknesses. At present Osmarender (via T at H) is more hackable and because distributed gives more immediately visible results. Mapnik has more unified design and deals neatly with some particular issues (text placement, bringing in and out data from other sources - eg coast, builtup areas at some zooms).
What I would like to see is the folks behind the metacarta version and those behind the Mapnik slippy implementation working together to see if one can inform the other. I am too naïve on the processes to offer solutions, but can whatever makes it posible for metacarta to render a planet dump quickly (massive server power?) be incorporated in some way in the process of updating the slippy version of the Mapnik layer (by port the tiles?)?

Cheers
STEVE

Steve Chilton, Learning Support Fellow
Learning and Technical Support Unit Manager
School of Health and Social Sciences
Middlesex University
phone/fax: 020 8411 5355
email: steve8 at mdx.ac.uk

Chair of the Society of Cartographers:
http://www.soc.org.uk/
Mind the (Map) Gap:
http://news.bbc.co.uk/1/hi/magazine/5413010.stm


-----Original Message-----
From: talk-bounces at openstreetmap.org [mailto:talk-bounces at openstreetmap.org] On Behalf Of Lars Aronsson
Sent: Thursday, May 03, 2007 6:40 AM
To: talk at openstreetmap.org
Subject: [OSM-talk] Mapnik has busstops



Yesterday on the #osm IRC channel, Ben exclaimed: Mapnik has 
busstops! And indeed, cute little buses show up as symbols on 
either side of the way, down at the deeper zoom levels.

Why had neither of us noticed this before?  Because when we get 
halfway down to that zoom level in the Mapnik layer, we're met 
with "more OSM coming soon", and then we back out and don't go in 
any deeper.  In fact, we don't go back to Mapnik at all.  Instead 
we use the Osmarender layer from Tiles at home, where the delay from 
edit to web presentation can be hours instead of weeks and months.

Now Tiles at home might have a melt-down over the full disk, but I 
guess that can be fixed in the next few days.  The only real 
threat to Tiles at home is if Mapnik would become fast.  And that's 
exactly what's happening. The Mapnik & tile engine by Schuyler 
Erle and Christopher Schmidt is updated on the same day that the 
new planet.osm becomes available, and tiles are generated on 
demand (and then cached) in a matter of seconds.  You just zoom 
in, and the tiles appear.  No more "coming soon"!  This 
presentation is available at http://labs.metacarta.com/osm/

This is very different from the Mapnik tile engine at 
www.openstreetmap.org, where requests for rendering are placed on 
demand, but nobody knows when these requests will be served.  
Come back an hour later or next week, and your new tiles might be 
there, or maybe not.

I don't know how it works.  Perhaps those who know can explain. 
All I know is that my new roads from the last weeks can be seen 
"in Mapnik" on labs.metacarta, but not yet on www.openstreetmap.

The Mapnik renderer (on both sites) uses some black magic for 
coastlines.  It doesn't read the coastlines that are now getting 
imported into the OSM database.  It provides correct coastlines at 
the outer zoom levels, but as you zoom in it shifts to another 
system that is completely broken.  Entire towns are found in the 
middle of a blue sea.  This bug in the Mapnik renderer might be 
the last hope for the Tiles at home subproject.  If Mapnik gets 
coastlines right and planet.osm is dumped on a daily basis rather 
than weekly, Tiles at home might go the way of the dinosaurs.


-- 
  Lars Aronsson (lars at aronsson.se)
  Aronsson Datateknik - http://aronsson.se

_______________________________________________
talk mailing list
talk at openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk