A document that carries other files inside it
ISO 19005-3 (PDF/A-3) exists because a document sometimes has to carry another file inside it, and that file has to come back out years later, whole.
Summary
An electronic invoice is the clearest case: one document carries the page a person reads and, inside it, the same amounts in a form accounting software reads without anybody retyping them. One file, one thing to keep, one thing to send.
It is not only invoices. A report can carry the spreadsheet its figures came from; a set of plans can carry the drawing they were produced from. The rule is the same each time: the carried file is named, described, dated, and listed where reading software will find it even knowing nothing of the archiving standard, so that opening the document is enough to get it back.
Carried is the word that matters. An archival document contains the files its reading depends on: they are in the document, not beside it on a disk. That is the whole promise of an archive — it is readable on its own, and it stays readable the day the folder around it is gone.
Technically
ISO 19005-3 (PDF/A-3) carries a file of any kind. ISO 19005-4 (PDF/A-4) carries one only if that file is itself archival; its F level (PDF/A-4F) takes any kind and asks for at least one — a document claiming PDF/A-4F with no carried file is refused.
Written around each carried file: the stream itself, its size, its date, the name in both the plain and the Unicode forms, the description, the stated relationship, the media type, and the entry in the name tree that makes ordinary reading software show it in its attachments panel. Under an archival claim, a file that no page, picture or drawing has claimed for itself is listed on the catalogue as associated with the document as a whole.
Stated by you: the bytes, the file name, the description, the media type, the relationship the file bears to the document, and the modification date. The date is not read off the clock — a document built twice from the same inputs comes out byte for byte the same, and a clock would break that.
Clause by clause
Beside a carried file the engine writes its size, its date and its media type, along with what you state. No checksum and no creation date are there: the standard does not ask for them, and the engine invents neither.
The engine does not check that a file carried under part 4 is itself an archival document. A document carrying anything else claims PDF/A-4F, and a checker is what says whether that was right.
For an invoice, the name declared in the document's description and the name of the file actually carried are not compared. Keeping them equal is yours, and it is what a receiving system matches on.
One rule here is the team's own and not the standard's. Both parts allow an archival document to hold an instruction opening a file that sits beside it rather than inside it — a neighbour on the same disk, or a page of another document. Under an archival claim the engine declines that instruction, and its refusal says which link or which field carried it: a document whose meaning depends on a neighbour stops meaning anything the day the neighbour moves, and an archive is the one document that has to be readable alone. It is the same line the engine draws on forms; the page on what reading software may settle for itself describes that. Outside an archival claim the instruction is written as asked.
What the engine does on its own
The engine does all of this as soon as a document claims the standard, whatever else the calling program asks for.
- Writes the carried file with its size, its date and its media type, and names it in both forms so all reading software agrees on the name.
- Registers it where ordinary reading software looks, so it shows in the attachments panel without anybody hunting.
- Lists it on the document as a whole, under an archival claim, when nothing else claims it.
- Refuses a document that claims PDF/A-4F and carries no file.
- Keeps everything the document leads to inside the document: an instruction opening a file beside it is refused under an archival claim, by the team's choice rather than by the standard's.
What you supply
The engine states what it was given and invents nothing else: without those values it refuses the document rather than choosing one for you.
- The bytes of the file, and the name it is to be found under.
- Its media type, and a description somebody reads.
- The relationship it bears to the document — source, alternative, data, supplement, or unspecified. It has no default: you state it for every carried file.
- Its modification date, stated rather than read off the clock.
The published texts behind these rules
| Published text | What it asks for |
|---|---|
ISO 19005-4 6.9 |
A document claiming PDF/A-4F names the files it carries. |
The other families of requirements
Which part of the archiving standard to claim Every font the document uses is embedded in the file Colours that mean the same thing on every device A page that comes out the same for everyone The document says what its marks stand for What the file says about itself Checking the claim, before and after