ISO PDF/UA and PDF/A in Hamburg
Martin Bailey at 16:52 GMT on 28 April 2009
Both PDF/UA and PDF/A got together as part of the ISO TC171 PDF committee meetings in Hamburg last week.
This was the first meeting of the PDF/UA (Universal Access) committee. Their scope is aimed at delivering a standard for PDF files that can be readily accessed by those with impaired vision or other disabilities.
As a first meeting, the first order of business was to select a convenor and a project lead. The secretariat for TC171 is AIIM, acting on behalf of ANSI, the American national standards body. Normal ISO practice is for the convenor and project lead to be from the same nation as the secretariat. Dennis Newman was appointed as convener, and Cherie Ekholm is now project lead.
We addressed all comments to the combined new work item/working draft ballot, including some interesting discussions as to whether it is necessary to ensure that a PDF/UA file can be formally validated as compliant. It seems clear that machine validation will not always be possible, as so many of the requirements are based on the intent of the content in ways that require a person to determine correctness. We agreed that it should be a goal that files & tools can be machine validated as much as possible, but that there’s no absolute requirement on machine validation. Human validation is sufficient to address the likely regulatory frameworks that may be built up around UA; machine validation would be a bonus.
The PDF/A committee (ISO TC171/SC2/WG5) met for two full days, and worked our way through 220+ comments on the last review draft of PDF/A-2 (ISO 19005-2). Fortunately a little over half of those were editorial or entirely non-controversial and could be disposed of very quickly.
Of the rest, the highlights were:
- Agreement that it’s very difficult to schema-validate XMP because different serialisations will result in failure to match. We discussed alternative methodologies, but it seems unlikely that those will bear fruit in the time-scale of PDF/A-2.
- PDF/A-1 allows the embedding of additional PDF/A-1 files within the main document. It had been suggested that we open up that restriction in PDF/A-2 to accommodate business needs for retention of machine readable (e.g. XML) data, original authoring formats (e.g. Excel documents) etc. After some debate it was agreed that this should be pushed out for consideration in PDF/A-3.
As the UK representative to these discussions I’d be very interested to hear input on this point, especially from UK organizations, both as to the need for support in PDF/A-3 and if anyone feels a real urgency for enabling embedding of non-standardized files within a PDF/A-2 file.
A small number of the issues were too complex and too technically detailed to be fully resolved in the full committee; those were delegated to an ad hoc committee comprising Leonard Rosenthol, Olaf Drümmer and me.
The next draft will include decisions made in Hamburg and those from the ad hoc team and should be out as a CD ballot by the end of June. The results of that ballot will be reviewed in the next PDF/A meeting in October in Orlando, FL, USA.
Some of the comments to the last PDF/A-2 review were put alongside input from elsewhere, including from the PDF/A Competence Center and together they add up to a sufficiently long list that we agreed to start work on a second technical corrigendum for PDF/A-1. If we decide to complete that task it could be published at about the same time as PDF/A-2.
Finally, we started discussions on the intended goals and benefits for a future PDF/A-3 standard. I’m sure those discussions will continue for some time to come!



