PDF/A-1
PDF/A-1 is ISO 19005-1, built on PDF 1.4, and the part with the narrowest feature set. It is also the part still named in most long-term archiving policies, so it is often the target even for documents that would fit a later part comfortably.
Two profiles cover it: pdfa.NewProfile1B for visual reproduction, and
pdfa.NewProfile1A for the same plus a tagged logical structure.
What PDF/A-1 forbids and the later parts allow:
| Feature | PDF/A-1 | PDF/A-2 and later |
|---|---|---|
| Transparency and transparency groups | Not permitted | Permitted |
| Optional content (layers) | Not permitted | Permitted with rules |
| Embedded files | Not permitted at all | Permitted |
| JPEG 2000 images | Not permitted | Permitted from part 2 |
| LZW compression | Not permitted | Not permitted |
| Reference and PostScript XObjects | Not permitted | Not permitted |
Because transparency is out, converting a modern document to PDF/A-1 is the conversion
most likely to change how the file looks. ApplyStandard flattens image soft masks and
removes transparency groups rather than failing, so the output is valid and the visual
result may not match the input. If the document uses transparency deliberately,
PDF/A-2 is the better target.
PDF/A-1 also carries syntactic checks that later parts dropped or relaxed: the
implementation limits in clause 6.1.12, the requirement that the document information
dictionary agrees with the XMP metadata, and the rule that a linearized file carries a
matching /ID in both trailers.
Where to look
| Guide | Covers |
|---|---|
| Apply PDF/A-1 | Converting a document to PDF/A-1, and what the conversion changes. |
| Validate PDF/A-1 | Checking a file against 1A and 1B, and reading the result. |
The shared mechanics, the full profile list and the conformance levels are covered on the PDF/A section page.