[Talk-GB] UK style rendering port ...
sk53.osm at gmail.com
Mon Jul 27 10:04:40 UTC 2015
Completely agree with Richard,
New rendering styles are a chance to start afresh with the technologies &
perspectives from 10 (or 11) years worth of experience. This does not just
mean vector tiles (although my feeling is these are still 'bleeding-edge'),
but also potentially more use of LUA and/or more complex database-side
post-processing: this is particularly true when considering rendering a
smaller area such as Britain & Ireland., where the constraints of database
update are less overwhelming.
There are plenty of trivial things which can be done now with a new render
which are harder for existing styles: such as making use of the hstore
A much tougher thing to try would be to find a way to include more of the
mapping community in establishing what is wanted: not just vociferous
voices on the mailing list & github, nor the bias inherent in a
On 27 July 2015 at 08:43, Richard Fairhurst <richard at systemed.net> wrote:
> Lester Caine wrote:
> > The other area that I am looking to roll back
> First, I wouldn't think in terms of "rolling back" at all.
> Cartography for a UK tileset could and should be designed from scratch. If
> you really want to start with an existing tileset then use OSM Bright, not
> osm-carto, but I'm not convinced that's necessary. We're much more likely
> get a coherent, UK-friendly style by thinking about it from first
> than by adapting osm-carto with its eight-year heritage (and all the
> attendant compromises).
> Second, let's not start with raster rendering. Let's build this as a vector
> tileset from the start, with in-browser/on-device rendering.
> Vectors have lighter infrastructure needs, and enable clickable POIs and
> on-the-fly restyling. The UK is small enough that we could just run it
> through tilemaker (https://github.com/systemed/tilemaker) without the need
> for a complex stack. The two disadvantages are that the very oldest
> will struggle, and that (for now) minutely updates don't happen unless you
> build that complex stack - but that's fine, we still have osm-carto for
> instant feedback.
> Let's build tomorrow's map, not yesterday's; and who knows, maybe osm-carto
> might want to borrow from some of our advances in months to come!
> View this message in context:
> Sent from the Great Britain mailing list archive at Nabble.com.
> Talk-GB mailing list
> Talk-GB at openstreetmap.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Talk-GB