[Mapcss] Zoom levels: bad idea?

Komяpa me at komzpa.net
Fri Jul 23 00:18:25 BST 2010


2010/7/20 Chris Browet <cbro at semperpax.com>:
> Unless I'm mistaken, I understand the zoom level (as in "way|z1-11") as "OSM
> TMS zoom level", right?
You may understand it that way if you want to.


> What about maps rendered in another projection than mercator/google, e.g.
> OSGB36, which is non-othogonal ?
They also have a bbox, and may have tiles (at least virtually), thus,
have zoom levels too.

> How is the implementation supposed to deduce the zoom level based upon a
> pixels per meters zoom level?
I guess defining that equator length on z0 is 256 pixels (.26 mm per
pixel) should be enough. Or, we may leave that up to implementer.

> Extreme, but I tend to think the zoom level namespace selector should be
> dropped altogether in favor of the "way|s5000" indicated as example in the
> specs, which is far more explicit.
That might be an alternative, but I'd prefer keeping zoom levels
as-is. If you look at most OSM stylesheets, they first do a conversion
of sacle into a zoom level and then use zoom levels as constants in
that case.


> While, if needed, going to zoom level from a scale in the context of a
> mercator projection is easy for the implementation, I don't see how to go
> from zoom level to scale consistently for "exotic" projections, given the
> restrictive (assumed) definition of the zoom levels...


-- 
Komяpa aka Darafei Praliaskouski
xmpp:me at komzpa.net mailto:me at komzpa.net




More information about the Mapcss mailing list