Collaborating in geospatial context since 2000!

Friday, July 30, 2010

(Geo)PDF is open and is more than graphics

I just got back from an interesting GeoWeb 2010 in Vancouver, BC. There I was asked why we picked PDF and not something "open" like SVG?

There are two problems with this question. I'll defer the "open" question and deal with the more insidious one: the conflation of PDF with simple graphics formats like SVG. By simple, I don't mean uncomplicated, just that they're only focused on one thing: graphics. PDF is better thought of as a document-centric container format, which includes in its definition a powerful page description syntax. It also defines ways of making annotations, adding interactive features, organizing pages and other document structure, embedding metadata related to the document, page contents, and almost any other element contained within a PDF file.

All of these elements are specified by an ISO standard: ISO 32000. Adobe made special arrangements with ISO to be allowed to freely distribute the draft specification that was adopted by ISO. You can find it on the all things PDF spec page, along with their latest extensions. Before it was a de jure open standard, it had always been a de facto open standard. Adobe made the spec open years ago. I've been referring to it since 2000. If you would like to read that sentence with a reverberating baritone "The Year 2000", feel free.

PDF is open under almost any reasonable definition of the word, I'd say.

TerraGo adopted PDF as the container format for GeoPDF products, because it *was* a de facto open standard for the delivery of interactive documents. By adding interactive features that guided people through a workflow, or allowed them to navigate through one more efficiently, TerraGo turned a collection of interactive document into an application.

When TerraGo creates a map book or other a GeoPDF compostiting application that brings down layers from ArcGIS Server and smushes it into a GeoPDF in the field, we're building applications -- not just pictures and images -- and we're building them on completely open standards, best practices, and published methods. No user or work-flow data is hidden in proprietary data structures in a GeoPDF product.

GeoPDF, while not a format, implies open. There is nothing to prevent someone from implementing software and applications that consume GeoPDF products and every byte of user data they contain without *any* reverse engineering. It's all right there!