The TEI customisation of the Welte-Mignon Advertisements Collection Encoding Pfeffer Niels Pfeffer Niels Tübingen 2026

Creative Commons Attribution 4.0 International License

Born digital.

 Written from a survey of the markup of the advertisements. The TEI customisation of the Welte-Mignon Advertisements Collection
What this customisation is for

The Welte-Mignon Advertisements Collection encodes print advertisements for the Welte-Mignon reproducing piano in TEI P5, one advertisement to a file (tei/ad*.xml); the encoding description that the files share is kept in tei/shared/encoding.xml. This ODD documents that encoding as a customisation of TEI P5, release 4.12.0. It admits only the elements that the edition uses, says what the edition means by an element or attribute where it gives it a meaning of its own, and closes the lists of values where the edition uses a closed set.

The customisation only restricts TEI: a file that is valid against it is valid against tei_all. Besides leaving out elements and closing lists of values, it requires that a person be linked to the GND and a place of publication to GeoNames by their URIs. Elements and values that a new advertisement needs are added here first; npm run survey-tei lists the elements, attributes and values that the files use.

The RELAX NG schema tei/schema/welte-ads.rng is generated from this ODD with the TEI Stylesheets of the same release (npm run build-schema in frontend/, see frontend/scripts/build-schema.sh). Every TEI file of the collection names that schema in its xml-model processing instruction and this ODD in the schemaRef of its encodingDesc. npm run validate-tei validates every file against the schema and, to make sure that the files remain conformant TEI, against tei_all of release 4.12.0. After changing this ODD, regenerate the schema and commit both; npm run check-schema tells whether they agree.

The schema

The schema takes its declarations from the TEI P5 source of release 4.12.0 (tei:4.12.0, that is https://www.tei-c.org/Vault/P5/4.12.0/xml/tei/odd/p5subset.xml). It includes the elements that the advertisements and the shared encoding description use, module by module, and documents the edition's use of elements and attributes where it is its own.

points to one or more topics, the categories of the taxonomy #topics in the classDecl: the recurring motifs of the advertisements that the passage (seg) or the picture (figure) shows. points to one or more renditions declared in the tagsDecl, each a CSS declaration of the typography of the original, e.g. #c (centred) or #uc (set in capitals). points to whoever is responsible for a topic, a description or a note: the editor (#np) or Claude, an AI model (#claude), as the titleStmt declares them. the main title: of the file, or of the advertisement or the periodical in a source the subtitle of the file: when the advertisement appeared the title of the advertisement, in a source the title of the periodical in which it appeared a short key for the firm; on the author in the titleStmt, the firm that placed the advertisement, by which the site groups the advertisements. describes one printing of the advertisement: its title (bibl), the periodical and the issue in which it appeared (biblFull) and the scan of the page (msDesc). A file has a sourceDesc for every printing found; decls on text points to the one that it transcribes. identifies the scan of a printing: the digital collection that holds it (repository) and the URL at which it shows the page (idno). the GeoNames URI of the place where the periodical appeared, e.g. https://www.geonames.org/2643743 (London). the date of the issue in which the printing appeared, as precisely as it is known the issue of the periodical, as it names itself the page or pages of the issue on which the advertisement appeared the URL at which the digital collection shows the scan of the page every rendition is a CSS declaration says whether the editor has checked the transcription and the encoding against the scan. checked by the editor not yet checked by the editor, e.g. an advertisement transcribed and encoded by Claude the page as the holding library serves it through IIIF; the first graphic of a surface, which has no type, is the scan on which the zones were drawn points to the sourceDesc of the printing that the file transcribes, the one on whose scan the zones of the facsimile were drawn. a region of the original page, which points to its zone in the facsimile with facs. a picture, described in a figure a part of a frame around the advertisement a block at the foot of the advertisement the brand name or a slogan, set off as a block of its own a panel that tells a story within the advertisement marks a part of the advertisement that is not transcribed: one that concerns other products than the Welte-Mignon, which a desc describes briefly, or one that cannot be read. concerns other products than the Welte-Mignon cannot be read on the scan a passage that shows one or more topics, to which ana points a slogan of the brand, e.g. The Master's Fingers on Your Piano an illustration of the advertisement, which a desc describes; ana points to the topics that it shows. a brand name, e.g. Welte-Mignon a performer, mostly a pianist, named as playing for the Welte-Mignon a composer, named for a work that the Welte-Mignon plays the URI of the person in the Integrated Authority File (GND) of the German National Library, e.g. https://d-nb.info/gnd/118697641 (Edvard Grieg). a testimonial of an artist or of another witness points to the sourceDesc of each printing that reads so.