[OSM-dev] osm pre-processing API. Was: Re: Intelligent fitting street names in OSMarender

[Robert Hart]

Barry,

I'm pleased to see that you are working on a pre-processing API. I wrote some of the Mercator projection code in osmarender, and did consider a perl pre-processing step at the time. In the end I went with the xslt version, simply because the advantages of pre-processing the projection did not justify new dependencies. Since then tiles at home has really taken off, and we're starting to see lots more really clever stuff.

To be honest I was fairly surprised you opted for a post-processing step for the bezierification, but I'm sure that's been discussed to death by now. The big advantage of doing it pre-processing is that we can put tags in the ways/nodes to fine-tune the process.

I don't think post-processing the text placement was necessarily a bad idea - it certainly needs to happen after the projection from geographic to image coordinates, and needs to know which render style is to be used in order to get fontsizes - but a pre-processing step isn't out of the question. 

I think the text placement hinting could easily be extended to check for overlapping text and repeated text (e.g. when both lanes of a dual carriageway are labelled).

Let me know when you have something setup and I'll update my code, and feel free to hack my code to death if you like. :-)

Rob


This message has been scanned for viruses by MailControl - www.mailcontrol.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20070310/24aa9831/attachment.html>