[OSM-dev] rendering: mad ideas

Immanuel Scholz immanuel.scholz at gmx.de
Fri Mar 24 15:29:15 GMT 2006


> I do not understand this discussion. Why should all programs that
> changes data in OSM be required to render tiles exactly in the same
> way as the applet renders them?  This seems to be a requirement that
> will hinder development of applications.

The idea is, that the applet users already have the rendered image created
on their machines within the applet. Now it is just a click with the
finger to upload this picture.

The discussion are also based on unverified assumptions:

- The server lacks massive on CPU power.
- Rendering images takes a relevant amount CPU power.
- Either server bandwidth is abundand or the amount of data is irrelevant
- People are willing to spend their bandwidth to upload rendered pictures

Personally I dislike the idea of having this in the applet. I would also
like to keep the rendering definition seperated from the API. *IF* the
above assumptions are really true, then a seperate client like osm at home is
not that hard to code (I would guess about two weekends hacking). And even
then, I really suggest choosing some protocol that keep the interfaces

Ciao, Imi.

More information about the dev mailing list