The set of media types specified by the Internet Assigned Numbers Authority.
The search phrase “IMT vs DCT” points at a real confusion in the DCMI namespace. Two terms look as if they do the same job: dcterms:IMT and dcterms:MediaType. They work at different levels, and once that is clear the pair becomes easy to use correctly.
A scheme, not a class of things
dcterms:IMT is a vocabulary encoding scheme. It does not describe any single format; it names a list, specifically the media types registered with the Internet Assigned Numbers Authority (IANA). Saying a value is “from IMT” tells a reader that audio/flac is meant as the registered media type and should be interpreted by the rules of that registry.
dcterms:MediaType, on the other hand, is a class. Its members are individual formats and physical media. audio/flac is one of them. So the relationship is membership: the media type is an instance of dcterms:MediaType, and it is a member of the scheme dcterms:IMT.
| Term | Type | Answers |
|---|---|---|
| dcterms:MediaType | class | “What kind of thing is this value?” |
| dcterms:IMT | vocabulary encoding scheme | “Which list was this value taken from?” |
| dcterms:format | property | “What is the format of this resource?” |
How IMT values look
The registry organizes types into top-level categories (application, audio, font, image, message, model, multipart, text, video) with a subtype after the slash. Some subtypes carry tree prefixes: vnd. for vendor types such as application/vnd.oasis.opendocument.text, and x- for unregistered experimental names, which the registration rules have long discouraged. Parameters can follow a semicolon, as in text/plain; charset=utf-8.
For metadata it is usually wise to record the bare type and subtype and keep parameters only when they carry real information, such as a character set for plain text. Type and subtype names are case-insensitive, and lowercase is the normal written form.
Declaring the scheme in practice
How the scheme is attached depends on the serialization. Qualified Dublin Core in XML uses xsi:type="dcterms:IMT" on a dcterms:format or dc:format element. In RDF the DCMI Abstract Model approach shown in the example builds a value node with dcam:memberOf and rdf:value. Linked data projects often skip that node and point dcterms:format straight at a URI that stands for the media type, which is simpler to query but leaves scheme membership implicit.
Whichever form is chosen, consistency across a collection is what lets an aggregator index the values reliably.
The registry also changes. New types are registered every year, and occasionally a long-used unofficial name receives an official one: text/javascript was confirmed as the standard type for JavaScript in RFC 9239 (2022), after years in which application/javascript was the recommended form. Records created under older advice are still correct for their time, but a normalization pass before publication keeps facets tidy.
Where IMT falls short
Media types were designed for transport: they tell a browser or mail client how to hand content to the right program. They were not built to distinguish format versions. application/pdf covers PDF 1.0 through PDF 2.0 and every archival and engineering profile in between. application/zip covers ordinary archives and also many container formats that registered no type of their own.
For that reason preservation systems pair IMT values with a more specific dcterms:FileFormat identifier, often a PRONOM PUID. The media type remains useful for delivery (it is what a server should send in the Content-Type header) while the finer identifier drives risk assessment and migration. The file format identification guide covers how to get both from a single tool run.
One last trap: file extensions are not media types. A file named .jpg may contain a PNG, and image/jpg is not registered; IANA lists image/jpeg.
Example in RDF
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:dcterms="http://purl.org/dc/terms/"
xmlns:dcam="http://purl.org/dc/dcam/">
<rdf:Description rdf:about="https://example.org/audio/interview-07.flac">
<dcterms:format>
<rdf:Description>
<dcam:memberOf rdf:resource="http://purl.org/dc/terms/IMT"/>
<rdf:value>audio/flac</rdf:value>
</rdf:Description>
</dcterms:format>
</rdf:Description>
</rdf:RDF>
Questions about dcterms:IMT
What does IMT stand for in Dublin Core?
IMT stands for Internet Media Type, the registry of type/subtype names such as text/html or application/pdf that IANA maintains. dcterms:IMT is the DCMI name for that whole list when it is used as a source of values.
What is the difference between dcterms:IMT and dcterms:MediaType?
dcterms:MediaType is a class: its members are individual formats or physical media. dcterms:IMT is a vocabulary encoding scheme: it identifies one particular list those members can come from. A value like video/mp4 is a MediaType that is a member of IMT.
Is IMT the same as a MIME type?
In everyday use, yes. MIME types and Internet Media Types refer to the same registry; the name changed as the types spread beyond e-mail to HTTP and other protocols. The registry lives at iana.org under media-types.
How do I state that a dc:format value comes from IMT?
In XML-based Dublin Core profiles the usual form is xsi:type="dcterms:IMT" on the format element. In RDF, the value node carries dcam:memberOf pointing to dcterms:IMT, with the media type string in rdf:value.
Is image/jpg a valid IMT value?
No. The registered type for JPEG images is image/jpeg. image/jpg appears often because some software emits it, but it is not in the IANA registry and should be normalized when records are cleaned.