Processing Digital Maps - Programming GPS and OpenStreetMap Applications with Java - page 86

with attributes to describe

ˆ roads: name, surface, width, lanes, accessibility, house numbering;

ˆ restrictions: gates, turns, times, vehicles, speed;

ˆ time dependencies: reversible one-ways, seasonal closure of roads;

ˆ signs: text, exit and route numbers;

ˆ multiple languages and voice phonemes (in various pronunciations)

and continually growing coverage (and information density) of America,
Australia, Europe, South Africa, Middle East, and other parts of the world.

7.6 Map Formats

The map specifications describe the content of a digital map. They do
not describe the proprietary structure of the company's map server, nor its
database. The internal database is optimized for concurrent map produc-
tion, eld worker's (audio) notes, and validation and must be exible for
modifications. Externally map suppliers should comply with international
standards to deliver navigable databases that can be used with third-party
tools. Different standards exist for different kinds of applications for desk-
top, server-based, or GIS navigation and routing. The formats differ with
dierent content, coding, and merging abilities depending on the customer's
business. After different portions of the map have been quality ensured,
they can be copied from the live database and extracted to widely available
data formats processed by system manufacturers.
One of the most prevalent formats in the industry is GDF 3.0, 5 GDF
incorporates a wide variety of features and is an international (slowly fading
out) standard for delivering navigable databases. Just like in OSM files,
GDF begins with a list of all nodes of the map (level 0). The information to
link the data is spread across the rest of the file using IDs as pointers. This
structure is extremely inecient for direct access by an application. Every
system uses different (map) features and every vendor uses a proprietary
data schema filtered and optimized for fast access (routing). The process of
converting map deliveries to a physical storage format of the target system
is called map compilation.
You will learn how to build your own map compiler to create dedicated
maps from openstreetmap for your own application (ROApp). The process
will give you a good idea of how time- and resource-consuming the process
is. Professional map compilers use the largest storage and fastest machines

5 GDF = Geographic Data File, based on CEN (European) Standard.