[OSM-talk] Norton, the village in the sea?

[SteveC]

* @ 10/05/06 09:40:38 AM ben at somethingmodern.com wrote:
> On Wed, May 10, 2006 at 09:07:12AM +0100, SteveC wrote:
> > * @ 09/05/06 11:30:36 PM openstreetmap-L at gj0.net wrote:
> > > Nick
> > > See 
> > > http://wiki.openstreetmap.org/index.php?title=Image:Yarmouth.png&rcid=6220
> > > 
> > > The label was placed deliberately, for rendering purposes,  so that it
> > > didn't obscure the streets.  It would make sense for the node to
> > > actually be in the place it is supposed to label.  Any ideas about a
> > > better way of doing this?
> 
> I know it sounds bizarre, but there's something subtle and interesting
> about a point in "real" space having the metadata "I'm a good place to
> refer to this other point."  Probably a good place for a key like
> "annotative at zoom level X".
> 
> > In a past life I did label placement and electrostatic-spring network
> > layout algorithms.
> > 
> > Put a spring with some spring constant between the labels and their
> > actual unmoving nodes. Then make all the labels repulsive with some
> > constant so that they move away from each other but cant stray from
> > their nodes because of the spring. Iterate the simulation with friction
> > and it tends to end up with nice label placement, after you play with
> > the two constants.
> > 
> > XSLT is turing complete right?
> 
> Oh good God.  Yes.  But anyone who can write a dynamic (read, iterative)
> Hooke's law sim in XSLT wins many-a-pint from me.

Yeah we need a ruby to XSLT compiler :-P

have fun,

SteveC steve at asklater.com http://www.asklater.com/steve/