A pre-existing related resource that is substantially the same as the described resource, but in another format.
This property is intended to be used with non-literal values. This property is an inverse property of Has Format.
Digital collections almost never hold a single file per object. A photograph has a lossless master, a compressed access copy and a thumbnail. A report exists as the author’s word processing file and the published PDF. A recorded lecture survives as a broadcast master, an MP4 and an audio-only MP3. dcterms:isFormatOf is the DCMI Metadata Terms property that ties those copies together, stating that the described resource is substantially the same as a pre-existing one, but in another format.
Three conditions in one definition
Each part of the definition does work.
Pre-existing. The target existed first. The property points from the derivative back to the resource it was made from: access copy to master, PDF to source document. In the example on this page the JPEG is the subject, and the TIFF it was generated from is the object.
Substantially the same. The content is unchanged in any way a user would care about. Resolution, compression or color space can differ; the picture, the text or the sound cannot.
Another format. The difference is encoding or carrier, which is what dcterms:format records on each resource. Two TIFF files at different resolutions are arguably the same format, and many collections still use the property for them, since the alternative properties fit worse.
The comment adds that values should be resources, not strings, and that the property is the inverse of dcterms:hasFormat.
Choosing among the relation properties
isFormatOf is one of several sub-properties of dcterms:relation, and the neighbors overlap.
| Situation | Property |
|---|---|
| Same content, different encoding | dcterms:isFormatOf |
| Revised, updated or adapted content | dcterms:isVersionOf |
| Content derived with changes (excerpt, transcript) | dcterms:source |
| Component of a larger resource | dcterms:isPartOf |
The boundary with dcterms:source is the one most often blurred. OCR text produced from a page image is derived from it, but it is not substantially the same: it loses layout and may contain recognition errors. A searchable PDF that embeds the page image and adds a hidden text layer sits closer to isFormatOf. Collections tend to settle such cases by local rule, which is fine as long as the rule is written down.
In preservation repositories
The property is common in repositories that keep a master and one or more derivatives per object. It lets a discovery system show the access copy while pointing archivists to the master, and it gives a migration process a way to find every derivative that must be regenerated when a master changes or is migrated to a new file format.
Some practical habits help:
- State both directions where the store allows it. Inference engines can derive
hasFormatfromisFormatOf, but many catalog systems are not inference engines. - Give every copy its own identifier and its own format statement. A derivative that inherits the master’s metadata wholesale ends up claiming to be a TIFF while being a JPEG.
- Point to the actual master, not to the abstract work. If the record describes an intellectual work, format relationships between files belong on the file-level records.
Values must be resources
Because the comment asks for non-literal values, isFormatOf "master.tif" with a plain filename is weak data: nothing can follow it, and filenames are rarely unique. A full URI under the repository’s own domain is the minimum. When files have no public addresses, internal URIs or URNs still work, provided they resolve inside the systems that read the records. In simple Dublin Core exports the relationship collapses into dc:relation, so the URI string is all that survives there, which is one more reason to make it a complete one.
Example in RDF
@prefix dcterms: <http://purl.org/dc/terms/> .
<https://example.org/objects/map-0193/access.jpg>
dcterms:format "image/jpeg" ;
dcterms:isFormatOf <https://example.org/objects/map-0193/master.tif> .
<https://example.org/objects/map-0193/master.tif>
dcterms:format "image/tiff" ;
dcterms:hasFormat <https://example.org/objects/map-0193/access.jpg> .
Questions about dcterms:isFormatOf
What does dcterms:isFormatOf mean?
It states that the described resource is substantially the same as another, pre-existing resource, only in a different format. A JPEG access copy made from a TIFF master, or a PDF exported from a word processing file, is a format of the original.
What is the inverse of dcterms:isFormatOf?
dcterms:hasFormat. If the access JPEG isFormatOf the TIFF master, the TIFF hasFormat the JPEG. Records can state either direction or both; the DCMI comment notes the inverse relationship, so consumers can infer one from the other.
When is dcterms:source better than isFormatOf?
When the content changed in the derivation. An excerpt, a transcript, a translation or a cropped detail is derived from the original but is not substantially the same resource. dcterms:source fits those; isFormatOf fits faithful format conversions.
Does the order of creation matter?
Yes. The definition says the related resource is pre-existing, so isFormatOf points from the newer resource to the older one. A preservation master created from an access copy would be unusual, but if it happened, the direction would follow creation order.
Is isFormatOf the same as isVersionOf?
No. isVersionOf covers a changed edition, adaptation or update of a resource. isFormatOf covers the same content in another encoding or carrier. A revised second edition is a version; the same edition as EPUB and PDF are formats.