Partial disclosure is the level of format documentation at which some, but not all, of the information needed to read and write a format correctly has been made publicly available.
Registries record the disclosure level of a format (full, partial or none) as one of the factors in assessing how sustainable it is.
A format can be perfectly well known and still not fully documented. Camera raw files are the usual example: most are built on the TIFF structure, so anyone can walk the directory of tags, yet the maker notes, the sensor calibration data and sometimes the compression of the image data itself are known only to the camera manufacturer. This in-between state is partial disclosure.
Format registries commonly grade disclosure on three levels. Full means a complete specification is public. Partial means some of it is. None means no official documentation exists, and everything known comes from the developer’s software or from reverse engineering.
Why disclosure is graded
The idea comes from sustainability assessment. The Library of Congress format description framework lists disclosure as the first of its sustainability factors, alongside adoption, transparency, self-documentation, external dependencies, impact of patents and technical protection mechanisms. The logic is simple: a format that can be read without the original software is more likely to remain readable, and documentation is what makes independent software possible.
Grading also says something that “proprietary” does not. Many proprietary formats are fully disclosed (PDF was an Adobe format with a public reference long before it became ISO 32000-1 in 2008). Some formats are developed openly yet documented only by their reference implementation. Disclosure measures documentation, not ownership or licensing.
What partial disclosure looks like
Partial disclosure takes a few recurring shapes.
- Documented container, undocumented content. Raw camera formats and some scientific instrument files follow a known structure but fill it with private blocks.
- Documented core, undocumented extensions. A base specification is public while vendor extensions (custom chunks, private tags, proprietary codecs inside a standard container) are not. A video file can be a standard container holding a stream whose codec is closed.
- Documentation under conditions. The specification exists but is available only to licensees or under a non-disclosure agreement. For an archive outside that circle, this counts as partial or none.
- Late disclosure. Microsoft’s binary Office formats (
.doc,.xls,.ppt) were widely used for years before Microsoft published detailed specifications in 2008. Files written before that date did not change, but the disclosure level of their format did, which is why some registries record when a disclosure claim became true (see effective claim date).
Practical consequences
For a repository, partial disclosure shifts risk onto the undocumented parts. A migration tool may read the documented structure correctly and silently drop the rest, so a converted raw file can look complete while losing the maker-note data that a later user needed. Two habits reduce the risk: keeping the original file alongside any migrated version, and recording which tool performed the migration and what it is known to ignore.
Partial disclosure can also change in the other direction. Vendor documentation that was once published can be withdrawn from a website, and a specification available only through a developer program disappears when the program ends. Keeping a copy of every disclosed document the repository relies on is cheap insurance.
Records should also cite what documentation does exist and say exactly what it covers, with a reference purpose that distinguishes official partial specifications from reverse-engineered notes. “Partial” with no explanation helps nobody. “Container documented in TIFF/EP; maker notes undocumented; decoding supported by LibRaw” gives a future reader something to work with. More on how documentation supports identification is in the file format identification guide, and related terms are listed in the vocabulary index.
Example in RDF
<format rdf:about="https://example.org/formats/vendor-raw-2009">
<name>Example camera raw format</name>
<disclosure>Partial disclosure</disclosure>
<note>Container follows TIFF/EP; maker note structure and sensor
calibration fields are not published.</note>
</format>
Questions about Partial disclosure
What does partial disclosure mean for a file format?
It means the owner or developer has published part of the information needed to work with the format. Typically the overall structure is documented while some components, such as proprietary metadata blocks, compression details or extensions, are not.
What are the other disclosure levels?
Full disclosure means a complete specification is publicly available, as with PNG or PDF since ISO 32000. No disclosure means nothing official is published, so knowledge of the format comes from reverse engineering or the developer's software alone.
Is a partially disclosed format a preservation risk?
Usually more than a fully disclosed one. Undocumented parts may hold content or settings that only the vendor's software interprets correctly. When that software is no longer available, those parts may become unreadable or be lost in migration.
Does disclosure mean the format is free to use?
No. Disclosure is about whether documentation is available. A published specification can still be covered by patents or license terms, and an openly licensed format can be poorly documented. Registries record these as separate factors.
Can reverse engineering replace disclosure?
Partly. Open-source projects have documented many closed formats in great detail. The result is useful but carries no guarantee: it describes the files the developers saw, and edge cases or later versions may behave differently.