GIOChannel* g_io_channel_unix_new           (int fd);
g_io_channel_set_encoding() function with NULL
GIOStatus   g_io_channel_set_encoding       (GIOChannel *channel,
                                             const gchar *encoding,
                                             GError **error);
guint       g_io_add_watch                  (GIOChannel *channel,
                                             GIOCondition condition,
                                             GIOFunc func,
                                             gpointer user_data);
gboolean    (*GIOFunc)                      (GIOChannel *source,
                                             GIOCondition condition,
                                             gpointer data);
	FALSE for end watch


Distance Point To Line:
(for Peuker)

d = |v^^.r| =   (|(x_2-x_1)(y_1-y_0)-(x_1-x_0)(y_2-y_1)|)
                /
                (sqrt((x_2-x_1)^2+(y_2-y_1)^2))

- Ausgleichskurve
	Ausgleichsrechnung
	Interpolation

* Peuker line simplification
* auto-split tracks
	if track has "gaps" then split track into two tracks.
	A gap is when in a group of four points A, B, C, D the distance
	from A to B and C to D is about the same but B to C is "a lot" more.	
* spike filter
	calculate mean deviation of all points of one track, then
	delete all points that have deviation > delta,
	recalculate mean deviation and redo
* mean of two tracks A, B
	for each point of track A calculate distance to
	track B and create a new point on a new track in the middle of the
	distance. Next do the same process for all point of track B.
	The new track lies in the middle between A and B.
	Requires component-wise distance function with direction, i.e. e.g.
	distance of point is x=-5 and y=3 to track.
* merge of two tracks
	for each point of track A find nearest corresponding point on
	track B, then calculate the average between both points and record
	this as new track point
	Repeat this for each point of track B.
	Limit the averaging to a maximum distance.


SQLite tables
-------------
tracks		track headers (streets, boundaries, rivers, ...)
	idx	pkey index
	tid	track id, id common to all data points of one track
	name	track-/streetname
	ttype	track-/streettype
	latmin	bbox latitude min
	lonmin
	latmax
	lonmax
trackpts	the actual track (all tracks combined in one table?)
	lat
	lon
	ele
	dtim
	nrsat
	hdop
	vdop
	pdop
points		POIs
cities		name, center location, "size"