Having problems with message search?
Btw, Leonard, can you give a brief summary as to the structure of the
PDF wrapper, at least how it compares with ZIP and a MIME-based
approach? In some ways, the PDF format had to consider issues similar
to what OR has to consider, such as user agent application access and
rendering of the internal content, including over networks.
Here is a VERY quick synopsis of my "PDF Internals" class that I offer to customers.
File Format
-----------
PDF files consists of four parts - Header, Body, XRef and Trailer (in that order).
- Header - identifies the file as being a PDF, and offers a CLUE to the
version
- Body - the "objects"
- XRef - the "object index", where to find each indirect object in the PDF
- Trailer - references to key "objects" on the PDF (root/Catalog, Encryption, etc.)
- Body - the "objects"
"Objects" in a PDF come in a variety (9) types - including simple scaler types (eg. Integer, Float, String), collection types (Dictionary and Array) and the Stream (for large data blocks). These base types are combined in pre-defined combinations (according to the PDF Reference) to make "higher order objects" such as Pages, Bookmarks, Annotations, etc. "Objects" can reference other objects - thus enabling a fine grained sharing/reuse mechanism.
A PDF producer isn't limited to the pre-defined "higher level objects" - you can always define your own (but not your own 'core types') AND you can also extend the pre-defined ones as well. For example, many PDF tools vendors add extra stuff to PDFs that they've processed so that what took place can be recovered later for "tracking purposes". This enables one to use PDF for all sorts of uses that Adobe never envisioned...
Compression/Encryption
----------------------
Compression, historically, was only allowed for data contained in a Stream object (page content, images, etc.) - but with Acrobat 6 (PDF 1.5) introduced compression of any collection of "objects" - thus enabling even smaller PDF documents (but breaking backwards compatiblity for the first time since Acrobat 2). Every major (and a few minor) lossless and lossy compression methods are supporting including Flate (ZIP), JPEG, CCITT G3/G4, JBIG2, and JPEG2000.
Encryption is supported in both symmetric (RC4 & AES) and asymmetric (RSA) varieties and can either be over all the objects or just a subset thereof (eg. encrypted PDF with plaintext metadata) It is usually accompanied by a set of "digital rights".
Digital Signatures are supported using standard RSA/X.509 technologies for either the entire PDF, one or more "byte ranges" within the document, or a collection of objects. DigSigs may be visible on the page(s) or invisible and may also include a set of "modification rights" which define what operations are safe to perform on the document w/o the signature being invalidated.
Content
-------
Each page of a PDF can have a series of "objects" associated with it that define what gets drawn on the page, using a syntax that is derived from (though different than!) Postscript. It is all about explicit placement of these objects on the page at defined coordinates. These content "objects" (which are different than the Body "objects") can be categorized into Text, Paths (vectors) and (Raster) Images. This is also a mechanism for reusable/shared content (eg. company logo, slide background).
Content can also be associated with "structure/tagging", which then applies semantic representations on the visual content. "This is a paragraph", "This is a Table Header", "This is a Figure", etc. This structure is useful for many things including enabling reflow, providing accessibility to the content, etc.
There are also a set of body "objects" that may appear (visually) on a page, but live on a special "layer" above standard content - this includes hyperlinks, markup & comment annotations, multimedia and Forms.
PDF also has standard "objects" for embedded files, sounds, movies/animations, etc.
So what does all this mean?!?
-----------------------------
It means that PDF is an excellent container for a variety of content, since it has native "objects" for handling the myriad of content that could be found in the documents of today - be they eBooks, journal articles, print publishing, etc.
It has pre-defined support for high quality text, raster and vector rendering along with markup & comment annotations. It has support for rich multimedia in a variety of formats. It supports structure/tagging to enable accessibility and even reflow.
It is also an open international standard with a VERY well deployed free reader already in existence and is the basis for other standards such as PDF/A. There are NUMEROUS open source (and commercial) libraries available for reading, writing and manipulating PDF documents in your choice of programming language (C/C++, .NET, Java, Perl, Python, Ruby, etc.) on your favorite OS platform(s).
How's that, Jon?
Leonard
P.S. Although I think that PDF would make an EXCELLENT "container" for OEP/OR, there are certainly valid arguments for the ZIP+XML solution as well.
---------------------------------------------------------------------------
Leonard
Rosenthol
<
mailto:leonardr@...>
Chief Technical
Officer
<
http://www.pdfsages.com>
PDF Sages,
Inc.
215-938-7080 (voice)
215-938-0880 (fax)
Offline 
Send Email