[OSM-talk] Too many nodes?

[Brian Quinion]

> > > This bezier stuff is nice for the renderers, but it shouldn't be a
> > > substitute for having enough nodes to define the curve. Don't forget
> > > are other data consumers/renderers that may not have the ability to do
> > > bezier curves. Extra nodes can always be reduced by using a
> > > algorithm if the number of nodes is a problem for a particular

> > While I agree with you what is 'enough'?  I'm doing hi resolution maps
> > for print - effectively zoom level 20 maybe 21 - and it would be
> > unreasonable to expect people to put in enough nodes to make that look
> > smooth!

> Of course, "enough" is subjective. I'm specifically thinking of maps on GPS
> devices. A certain amount of jaggedness is expected, but it should be a
> reasonable approximation of the curve.

Sorry - I think my previous email came across badly... anyway:

You make a good point - I was thinking of the number of points in
terms of the quality of the output I want but it probably makes more
sense to think in terms of the reliability of the input data.  A
guideline of 'the limit of the quality of your GPS / data source'
sounds like very good advice.

It comes down to that fundamental conflict of the whole OSM project -
what we capture is data but what most of us want out of the end is
good quality maps!

--
 Brian