Sustainability of Digital Formats
 Planning for Library of Congress Collections

ExtensionLevel 3 of PDF (Portable Document Format), introduced by Adobe Systems with Acrobat 9.0 in June 2008, included a number of new features, including "a geospatial coordinate system" and a "number of PDF constructs ... to support geospatially registered content." The important constructs include two dictionaries for a viewport (a region on a page, assumed to contain an image).

The first geospatial construct is a new GEO subtype of measure dictionary. This type of measure dictionary defines the relationship between points or regions in the two-dimensional viewport space and points or regions with respect to an underlying model of the earth. A geospatial measure dictionary describes the bounds of the map in relation to the rectangular viewport (also known as a neatline), specifies a projected or geographic coordinate reference system that applies to the source data, and optionally provides a different coordinate system to use for display and preferred display units. The coordinate systems are defined either by using EPSG codes or by using WKT (well known text). See Notes below for more details on EPSG codes and WKT, the two de facto standards for referring to pre-defined coordinate systems. The geospatial measure dictionary must also have an array of registration points used to tie points on the map image to geospatial points defined in the coordinate space.

The second geospatial structure is a new dictionary type, PtData. A PtData dictionary comprises two related arrays. The syntax and semantics are specified in an array of names for data elements with predefined names for three elements: LAT, latitude in degrees; LON, longitude in degrees; ALT, altitude in meters. The second array XPTS is an array of arrays corresponding to the named elements. This structure is designed to allow attribute values to be associated with points and presented to users within a PDF viewer. It highlights the fact that Adobe's geospatial features for PDF are intended to support 3D geospatial data as well as 2D.

In Adobe Supplement to the ISO 32000, BaseVersion:1.7, ExtensionLevel:3, available at http://www.adobe.com/content/dam/Adobe/en/devnet/acrobat/pdfs/adobe_supplement_iso32000.pdf

For encoding geographic coordinate systems in the measure dictionary, either of two mechanisms may be used. In either case, the objective is to specify unambiguously the geodetic datum, geoid/ellipsoid/spheroid, coordinate system, and any map projection in use for the image or data.

• EPSG (European Petroleum Survey Group) codes. These are assigned by the Surveying and Positioning Committee of the International Association of Oil & Gas producers (OGP), formerly known as the European Petroleum Survey Group, and made available as the EPSG Geodetic Parameter Dataset at http://www.epsg-registry.org/. From Using the EPSG Geodetic Parameter Dataset: "The EPSG Dataset is a repository of parameters required to: define a coordinate reference system (CRS), which ensures that coordinates describe position unambiguously; and define transformations and conversions that allow coordinates to be changed from one CRS to another CRS. As an example, 4269 is the EPSG code for the coordinate reference system usually referred to as "NAD83".
• WKT (well known text). WKT values are structured, human-readable, textual strings, that describe the elements of a coordinate reference system. They have been declared by the Open Geospatial Consortium and documented in an appendix to OGC 06-103r4: OpenGIS Implementation Specification for Geographic Information - Simple feature access - Part 1: Common Architecture. Essentially the same document has been published as ISO 19125-1:2004. WKT is employed in many GIS systems. The WKT equivalent of EPSG code 4269 is:

  GEOGCS["GCS_North_American_1983",
  DATUM["D_North_American_1983",
  SPHEROID["GRS_1980",6378137.0,298.257222101]],
  PRIMEM["Greenwich",0.0],
  UNIT["Degree",0.0174532925199433]]