What exactly is this PDF problem? Readers who have exchanged PDF files with clients for years without ever running into problems may be forgiven for wondering what this supposed PDF problem is. The problem arises because there are at least two ways to create PDF files.

One way to create a PDF file is by placing a paper original of a document into a sheet scanner, and optically scanning the paper original. The light and dark areas on the paper original are scanned and the PDF file contains ones and zeros indicative of which areas were light and which areas were dark. One name for such a PDF file is a "bit-mapped raster-scanned" PDF file. At the receiver's computer, the file is displayed or printed simply by displaying or printing light and dark areas that match the areas that were light or dark in the paper original that was optically scanned.

A second way to create a PDF file is by "printing" from a word processor or other character-based application into a PDF file using a PDF print driver. Such a PDF file contains bytes indicative of particular characters that are meant to appear on the later-rendered page, as well as bytes indicative of the font and point size to be used in the later rendering and the position of the characters on the page when rendered. One name for such a PDF file is a "character-based" PDF file. At the receiver's computer, any display or printing of the PDF file requires that the receiver's computer possess font information for the fonts that were used. Usually this is not a problem because most character-based PDF files use fonts that everybody has on their computers.

A particular PDF file might be a hybrid of these two types, with some pages created by optical scanning and other pages created by character "printing" by means of a driver. For example Adobe Acrobat (the full version, not the reader) may be used to stitch together pages from two or more PDF files into a single PDF file. The source PDF files might have been created by different means, e.g. one by scanning and one by character printing.

Experience shows that every now and then, a font installed on an author's computer will either fail to match a seemingly identical font on the receiver's computer, or will be missing from the receiver's computer. Adobe's Acrobat documentation says:

If you do not embed fonts in a PDF file and a user opens the file on a system that does not have the file’s fonts, Acrobat temporarily substitutes fonts. For Roman text, Acrobat uses serif and sans serif Multiple Master fonts to simulate the original font. For Asian text, Acrobat uses fonts from the installed Asian Language Kit or from similar fonts on the user’s system.

The result of such substitution can be a received PDF file which, when rendered on the receiver's computer, has blanks where some characters should appear. Or the result can be a received PDF file which, when rendered, has page breaks in unexpected places.

One way to avoid completely any risk of non-identical rendering at the receiver's computer is for the author to avoid character "printing" and instead to use only optical scanning of paper originals in the creation of PDF files. In a character-based PDF file, one way to reduce the risk of non-identical rendering at the receiver's computer is to configure the PDF driver so that it will "embed" any fonts used within the PDF file. This makes the PDF file larger but permits the receiver's computer to render such characters even if their font had not previously been installed on the receiver's computer. Yet another approach would be to design and use a PDF printer driver that would never store character-based information at all in the PDF file, but would instead store only rendered bitmaps in the PDF file, even for files printed from character-based applications.

Still another approach would be to avoid the use of PDFs in e-filings, and instead to use (for example) multipage TIF images, which are always (by definition) raster-scanned bitmaps, and which never (by definition) contain character-based information.

Still another potential complication arises because some fonts cannot be embedded in PDF files. The Acrobat documentation says:

A TrueType font can contain a setting added by the font’s designer that prevents the font from being embedded in PDF files.

Some designers of fonts assert proprietary rights in their fonts, demand license fees for use of their fonts, and object to their fonts being embedded in PDF files. This puts a patent office in the position of wondering how to handle a PDF file that relies upon a "licensed font" which the patent office does not possess or has not licensed.

The above discussion merely describes briefly the areas of concern that surround PDF as a format for e-filed documents. The subject of PDF portability is a complex one. Yet the plain fact is that lots of people send PDF files to each other without encountering any of the above-mentioned problems. More experience will be needed to assess just how often these problems really arise.