[OSM-dev] Converting OSM Mapnik stylesheet to Cascadenik or Carto

[Andy Allan]

On Thu, Jul 28, 2011 at 12:34 AM, Frederik Ramm <frederik at remote.org> wrote:
> Hi,
>
> Graham Jones wrote:
>>
>> I picked Carto over Cascadenik for the same reason - that it seems to be
>> developing actively, and is reported to be significantly faster.
>
> Has anyone done a proper comparison between MapCSS and Carto? Does Carto
> have any advantages?

No comparison that I'm aware of, but carto certainly has advantages. I
would say that MapCSS is very primitive when you consider the range of
symbolizers and attributes that mapnik (and hence cascadenik/carto)
support compared to any of the engines that use MapCSS. Also MapCSS is
(somewhat pointlessly) tied up dreadfully into OSM syntax with
selectors based on not-quite-OSM-primitives like "relation" and "way",
as well as suffering hugely from the NIH syndrome when it comes to
having named every single attribute just slightly differently from
those which were already available in the mapnik/cascadenik world, for
no fathomable gain.

The suggestion to "fix tilemill to support MapCSS like everyone else
does" is laughable, really. I would certainly promote convergence in
the syntaxes, but that's more likely to be by MapCSS tacking towards
everyone else than the other way around.

But there is a fundamental issue with resolving the two approaches, in
that in the "mapnik world" they are rendering more-or-less Simple
Geometries, so you need to pre-process OSM primitives into point,
lines, polygons. MapCSS takes the (misguided, imho, see my SOTM-EU
talk) aim to support rendering OSM data directly, so it's full of
nitty gritty OSM-specific data processing in the syntax. I've no idea
what would be needed to hook that up to something like pgsql+mapnik.

Cheers,
Andy