PDF/A – Comparing Subsets

by Dr. Uwe Wächter

Our customers often ask us, “Which should I be making – PDF/A-1 or 2?” Today we’re going to discuss this question.

The ISO standard PDF/A-1 was published in 2005. The goal was the targeted restriction of the PDF’s scope of function, in order to make standard-conforming files archivable long-term. PDF was still defined over an Adobe reference at the time. With PDF/A, no new PDF reference could be created. Instead, recommendations were marked as obligatory and ambiguities from limitations clarified, e.g. transparent objects were forbidden.

PDF/A has the following requirements:

  • Based on PDF reference 1.4
  • All fonts must be embedded
  • Colors must be specified
  • Forms, commentary, notes partially permitted
  • No multimedia objects included
  • No embedded files
  • Compression for selected objects only
  • Compression: no LZW or JPEG2000
  • No encryption
  • No transparent objects
  • No optional content
  • Rules for metadata on the basis of XMP
  • PDF/A must be identified as such

 

In 2011, Part 2 of the standard was published: PDF/A-2.

The most important innovation was certainly the fact that this version can now build on PDF as ISO standard, ISO 32000-1 (corresponds to PDF 1.7). In the meantime, Adobe’s definition, implicit komposition3Definition with the prevailing XMP, became ISO standard as well. Through the further development of the PDF format, newer functions can also be permitted for PDF/A. This includes for example the OpenType fonts, JPEG2000 image compression and newer types of commentary.

In the original PDF reference, transparencies were described somewhat vaguely. In the PDF ISO standard, they are now so precise that they can be safely used in long-term PDF. In other words, PDF/A-2 permits transparencies. So you no longer have to worry about PDF/A generations.

PDF/A-1 had, on the basis of PDF 1.4, a size limitation of 200 inches, or 5.08 m, which was too small for many applications. The work-around was a scaling in the generation of a PDF/A. In the PDF/A file itself, however, the original size was then no longer recognized. PDF/A-2  now has a standard built-in scaling factor, the user units. With it, extensions of around 300 km are now possible.

With PDF/A, one often missed PDF’s usefulness of layers. This PDF function is now likewise better standardized: one layer must be designated as default layer. Then you can actually use layers which, for example, are handy in the conversion from CAD models.

PDF/A-1 made no requirements for digital signatures, and as a result these were always permitted. As PDF innately allows for manageable embedded signatures, this function was also popular. But PDF/A-2 is the first version which contains the reference to signature standards, so that there are no errors in the use of signatures for long-term archiving.

Another important innovation was the possibility for embedding other files. If these files again are PDF/A-conform themselves, this feature can be used.

 

PDF_doppeltIn 2013, Part 3 was published: PDF/A-3.

The only new feature here is the ability to embed arbitrary file formats. Just one identification must be made, in how the embedded files stand in relation to the original PDF/A-3: previous version or other representation.

PDF/A-3 makes no guarantees regarding the quality of the embedded files. In addition you must impose rules yourself, so as not to proliferate. For the possibility to package any files in PDF/A-3, there are numerous applications:

  • 2D PDF views of a design with stored original 3D geometry
  • invoice as PDF and stored invoice data as XML (ZUGFeRD)
  • original existing non-PDF/A file but with valid digital signature embedded in the PDF/A file derived from it.

 

And what are levels all about?

PDF/A-1 comes with the levels a and b. For PDF/A-2, there is also level u.

Level b is the lowest, but guarantees the PDF’s reliable visual reproduction. This level is usually sufficient. With the sources available it is often the case that no higher level can be reached.

Level u also demands that all characters in the file have Unicode mapping. Mind you, if you need Unicode, you can do that with level b. You just have to use only one Unicode font.

Level a is the complete standard. This includes demands for automatic accessibility. Objects like headers, images, or indexes must be identified as such. This is the first step toward PDF/UA.

Have any questions? Talk to us!

Dr. Uwe Wächter
uwe.waechter@sealsystems.de

 

Share

Leave a Reply

Your email address will not be published. Required fields are marked *

* = Required field

  • Blog Categories

  • Next Events

    1. DSAG Technologietage

      20. February 2018 - 21. February 2018
    2. Siemens PLM Connection

      19. June 2018 - 20. June 2018
    3. DSAG Congress

      16. October 2018 - 18. October 2018
    4. PLM Europe – Siemens PLM Connection 2018

      29. October 2018 - 31. October 2018