A domain facet type is a controlled category that groups file formats by the broad kind of content they represent, such as audio, image, text, video, spreadsheet or database.
Registries use domain facets for browsing, filtering and policy, and allow a format to carry more than one facet when its content spans domains.
Browse any format registry and the first filter on offer is usually by kind: audio, image, text, video. That first cut is a domain facet. It answers the most basic question a user has about an unfamiliar format, “what is in it?”, before any question about versions, signatures or compression.
Facets, not a tree
The word facet comes from library classification, where S. R. Ranganathan’s Colon Classification of the 1930s described subjects by combining independent aspects instead of placing each subject at one point in a hierarchy. Format description benefits from the same approach. A format has a domain, and separately it has a text or binary form, a compression type, a byte order and a disclosure level. Treating domain as one facet among several keeps each axis clean.
It also allows several values on the same axis. SVG is a vector image and an XML document. An EPUB file is a text format and an aggregate (a ZIP container with a defined structure). A database dump can be text and database at once. A strict tree would force a choice; a facet scheme records both.
Typical facet values
Registries define their own lists, and PRONOM’s format types give a good sense of the usual range, with values such as “Image (Raster)”, “Image (Vector)”, “Text (Mark-up)”, “Audio”, “Video”, “Spreadsheet”, “Database” and “Aggregate”. A broadly comparable set:
| Domain | Examples |
|---|---|
| Audio | WAV, FLAC, MP3, AIFF |
| Video | Matroska, MP4, AVI |
| Image (raster) | TIFF, PNG, JPEG |
| Image (vector) | SVG, EPS |
| Text | plain text, XML, HTML, source code |
| Spreadsheet / database | XLSX, ODS, SQLite, dBase |
| Page description | PDF, PostScript |
| Aggregate | ZIP, TAR, disk images |
| Geospatial / 3D / font | shapefile, glTF, OpenType |
A domain is usually assigned at the level of the format family and inherited by its versions, since a new TIFF revision does not stop being a raster image format. Exceptions exist: a later version can add a capability (vector layers in an image format, embedded media in a document format) that justifies a second facet only from that version on.
The categories are not perfectly parallel. “Page description” is defined by function, “aggregate” by structure, “audio” by content. That unevenness is normal and tolerable, as long as every value has a short definition so catalogers apply it the same way.
What domain facets are used for
Navigation. Users who do not know a format’s name can still find it under its domain. A researcher holding an unknown file with a .dat extension can at least narrow the search once a quick look shows it contains sampled sound or pixel rows.
Policy. Preservation plans are often written per domain: master formats for raster images, normalization rules for audio, significant properties for spreadsheets. The compression lossiness that is acceptable, for instance, differs between image and text domains, because a text format has no tolerable loss at all.
Identification triage. When a tool returns several candidate formats for one file, the domain of each candidate is a fast plausibility check. A file found in a folder of field recordings that matches both an audio format and a spreadsheet format is almost certainly the former, and the domain facet makes that reasoning explicit.
Reporting. Repository statistics such as “how much audio does the repository hold?” depend on consistent domain tagging of every identified format.
Domain facet versus resource type
The DCMI Type Vocabulary used with dc:type looks similar (Image, Sound, Text, Dataset, MovingImage), but it describes a resource, not a format. A digitized photograph is an Image; its TIFF file belongs to the raster image domain. Most of the time the two agree. They part ways for container and aggregate formats: a ZIP holding a thousand photographs is a collection of Images, while its format domain is aggregate. Mapping tables between the two schemes help, but each record should keep both values. More on identification and grouping is in the file format identification guide and the vocabulary index.
Example in RDF
@prefix ex: <https://example.org/vocab#> .
<https://example.org/formats/svg-1.1>
ex:name "Scalable Vector Graphics 1.1" ;
ex:domain ex:ImageVector , ex:TextMarkup .
<https://example.org/formats/flac>
ex:name "Free Lossless Audio Codec" ;
ex:domain ex:Audio .
Questions about DomainFacetType
What is a domain facet in a format registry?
It is a category that says what kind of content a format holds: sound, still images, moving images, text, tabular data and so on. It is called a facet because it is one independent axis of classification among several.
Can a format belong to more than one domain?
Yes, and many do. SVG is both a vector image format and an XML markup format. PDF can be a page description, an image container and a document with embedded data. A facet scheme allows several values per format.
Which domain categories are typical?
Common ones include audio, video, raster image, vector image, text (unstructured, structured or markup), spreadsheet, database, presentation, page description, geospatial, 3D model, email, font and aggregate or archive formats such as ZIP and TAR.
How is a domain facet different from dc:type?
dc:type describes the nature of one resource, often with the DCMI Type Vocabulary (Image, Sound, Text, Dataset). A domain facet describes a format, which is a class of resources. The values look alike, and mappings between them are common.
What are domain facets used for?
Browsing a registry, filtering identification results, and setting policy. A repository might apply one preservation plan to all raster image formats and another to audio, which only works if every format is tagged with its domain.