Cataloguing notes
On saving, a MARC21 record does not remain an isolated bibliographic record: an automatic cascade reads it field by field and generates/links the entities of the IFLA LRM model (Work → Expression → Manifestation → Item), then publishes them as RDF triples on a server suited to hosting the converted data. This page explains, field by field, what is read from the MARC21 and what it generates on the other side — so that cataloguers can understand why a field must be filled in a certain way in order to obtain correct linked data (Linked Open Data).
The cataloguer describes, the system connects. A URI (the unique address that makes a piece of data "linked") is never invented from the free text of a field: it is always taken from a dedicated subfield (almost always $1, more rarely $0) pointing to an authority already existing in the catalogue (an Agent, a Work, a Place, a Subject...). If that subfield is missing, the data remains descriptive text in the MARC but does not enter the network of linked data. In this way the system stays faithful to what the cataloguer chooses to describe.
Beyond what is written in the MARC, if a linked agent, place or subject already has a Wikidata identifier known to the catalogue, the system automatically completes, only for fields still empty, without ever overwriting data already present: VIAF/LOC/ISNI/BnF/SBN identifiers, an agent's place and year of birth/death, a place's coordinates and country, a subject's BNCF identifier. If Wikidata does not respond or the agent has no QID, simply nothing happens — saving the record is never blocked by this step.
required for LOD = without this subfield the entity is not linked at all (it remains text in the MARC only).
creates/links automatically = the system creates the entity on the fly if it does not already exist (no extra manual action required).
enriches = optional field that adds a piece of data but is not essential to connecting the main entity.
The abstract entity: the intellectual or artistic content itself, independent of any language, translation or edition — e.g. "Sophocles' Antigone", not any particular translation or reprint of it.
| Field | What it represents | What it generates/links on the LOD side | |
|---|---|---|---|
| 100 $1 | URI of the creating agent, chosen from the authors dropdown | Links the Work to its creator (rdaw:P10061). The agent must already exist as an authority — it is never created on the fly from here. |
required |
| 100 $a, $d, $e | Name, dates, qualifier of the creator (textual form) | They remain in the MARC as human-readable text; they do not generate a URI of their own (the URI comes from $1). |
descriptive |
| 240 / 130 $a | Uniform title (original title of the Work, for a translated work) | Heading of the Work. If absent, the manifestation title is used (245 $a) — correct only for works that are NOT translations. | creates/reuses |
| 240 $f | Date of composition/first publication of the original work | Time-span of the Work (LRM-R9). It also accepts non-numeric text ("4th century BC", periods) as well as a plain year. | creates/reuses |
| 240 / 130 $1 | URI of the Work, if the record is already linked to an existing Work | Reuses the Work instead of creating a new one (idempotent cascade); written automatically by the system itself after the first save. | automatic |
| 245 $a (riserva) | Title of the publication, used only if 240/130 are missing | Sets the Work's default preferred title — a starting value that often needs manual correction when the Work is a translation (the published title doesn't match the original one). | automatic |
| 041 $h | Original language of the work (if different from $a, i.e. for a translation) | Language of the Work — distinguishes the original Work from its translation. If the 245 indicates a bilingual ("facing text") edition, it also generates a second Expression in the original language, linked to the same Manifestation. | for translations |
| 246 $a | Varying form of title (e.g. title in another language, cover title) | Generates a variant Nomen (LRM-E9) linked to the Work. Unlike Subjects and Agents, it is never automatically removed if the 246 disappears in a later save. | creates/links |
| 650 $a, $x, $2 | Subject, subdivision, source of the terminology | Subject created if it does not exist, linked to the Work. The link is synchronised in both directions: if a 650 is removed in a later save, the link generated from the MARC is removed automatically (subjects added by hand from the LRM form are never touched). The same mechanism is also used for agents linked via 700 $1. | creates/links/syncs |
| 700 $1, $4 (coautore/curatore) | A co-author (relator "aut"), or an editor (relator "edt") when the Manifestation has no 100 at all | Links the agent as a co-creator of the Work (LRM-R5). The editor case models an aggregating Work such as an anthology: whoever edits the whole collection is treated as a creator at the Work level, not just responsible for this single Manifestation. | creates/links/syncs |
| 730 $1 | URI of a related work contained in the same publication | Work↔Work relationship of type "related work" (LRM-R1) — used for anthologies/collections. | links |
| 773 / 800 / 810 / 811 / 830 $1 | URI of the "parent" work (bibliographic series, multipart work) | Work↔Work relationship of type "part of"/"has part" (LRM-R1), in both directions. | links |
The intellectual or linguistic realization of the Work: a text in a specific language, in a specific form — e.g. "Giovanni Greco's Italian translation" of Antigone.
| Field | What it represents | What it generates/links on the LOD side | |
|---|---|---|---|
| 041 $a | Language of the published text | Language of the Expression. | required |
| 041 $h (confronto) | Original language of the Work (041 $h), compared with this Expression's language (041 $a) | If the two languages differ, the system recognizes a translation and links this Expression to the original-language Expression of the same Work (LRM-R24), if already cataloged. The link is created only when the Expression is first created, not recalculated on later saves. | automatic |
| 245 $a, $b ("a fronte") | The phrase itself in the title/subtitle, e.g. "with facing Latin text" | If detected, and the original language (041 $h) differs from the text's language, creates or reuses a second "parallel" Expression in the original language, linked to the same Manifestation — for bilingual facing-page editions. | creates/links |
| 245 $a + resp. + lingua | Title, contributors and language already resolved from the fields above | Builds the Expression's descriptive label (e.g. "Antigone. Italian (Greco, Giovanni)") — recalculated automatically on every save, never editable by hand. | automatic |
| 700 $1, $4 (trad./cur./collab.) | URI and role of the agent responsible for this specific Expression (translator, editor, contributor) | Links the agent to this Expression — synced on every save, like the Work's Subjects: an agent removed from the 700 is automatically unlinked. | creates/links/syncs |
| 700 $t, $l | Title and language of a related work (for anthologies/aggregating works) | If the title matches a work linked via 730, it creates the specific Expression of that work with this agent as translator/editor. | aggregates only |
The concrete editorial embodiment of the Expression: this particular printed edition, with its publisher, year, format — e.g. "the 1982 Einaudi edition" of that translation.
| Field | What it represents | What it generates/links on the LOD side | |
|---|---|---|---|
| 020 $a | ISBN | Identifier of the Manifestation — also used to determine whether a Manifestation already exists (avoiding duplicates) and for MARC/RDA deduplication in search results. | automatic |
| 245 $a, $b, $c | Title, subtitle and statement of responsibility of the published manifestation | Title of the Manifestation (cleaned of trailing ISBD punctuation). The 245 $c is also parsed to detect a "parallel text" edition. | required |
| 264 $a, $b, $c | Place, publisher, year of publication (text) | Descriptive text of the Manifestation + Time-span of the year of publication. | required |
| 264 $0 | URI of the place of publication, chosen from the dropdown or from a GeoNames search | Links the Manifestation to the Place as an authority, not merely as free text. | creates/links |
| 264 $1 | URI of the publisher, chosen from the corporate bodies dropdown | Links the Manifestation to the Publisher as an authority, with the publisher role. | creates/links |
| 300 $a, $c | Extent (pages and/or number of physical volumes) and dimensions | Pages of the Manifestation; if it indicates several volumes ("2 v."), it determines how many Items (copies) are generated. The $c (e.g. "22 cm") goes into the Manifestation's physical dimensions. | automatic |
| 490 $a, $v | Title and numbering of the publisher's series | Series created on the fly if it does not exist, linked to the Manifestation. | creates/links |
| Leader pos. 6 + campo 007 | Record type and physical carrier characteristics | Determines the Manifestation's media/carrier type (RDA 337/338 — e.g. printed text, electronic resource, microform, sound recording). Computed automatically; if Leader and 007 aren't recognized, the default stays "printed text". | automatic |
| 008 (pos. 07-10, 35-37) | Fixed field: publication date and language, encoded by position rather than in subfields | Automatic fallback for the Manifestation's publication year (if 264 $c is missing) and for the Work's original language (if 041 $h is missing) — no action needed. | automatic |
| 100 $1 (ancora strutturale) | The same creator agent already linked to the Work (see above) | Also recorded as the Manifestation's author, but only when it is first created — a structural "anchor" never touched again on later saves, unlike the other contributors (700) which stay synced every time. | automatic |
| 700 $1 (tutti i ruoli) + 264 $1 | Every responsibility agent listed in the record, including any publisher (264 $1) | Transcribes onto the Manifestation ALL the record's contributors — not just the ones relevant to the Work or the Expression — synced on every save: an agent removed from a field is automatically unlinked here too. | creates/links/syncs |
| 700 $4 | MARC relator code (URI of id.loc.gov/vocabulary/relators/...) |
Determines the role: aut→author, trl→translator, edt/com→editor, pbl→publisher, aui/ill/ann→contributor. |
required |
The single physical copy owned by the library: the concrete object on the shelf, with its own shelf mark and condition.
| Field | What it represents | What it generates/links on the LOD side | |
|---|---|---|---|
| 300 $a (riletto) | Number of physical volumes, if stated (e.g. "2 v.") | Determines how many Items are generated — one per volume, each with a "Vol. X of Y" note — instead of a single generic Item. | automatic |
| 100 + 650 (derivato) | No dedicated MARC field: computed from the same data already used for the Work and Subjects | Generates the shelf mark (e.g. "Narrativa.GRE-01") and assigns a library section based on the Work's linked subjects (default section: Saggistica/Essays, if no subject suggests otherwise). | automatic |
Some Item fields are not derived from MARC at all: the library (always "MyLODLibrary"), availability status (always "available"), and the cataloger who created it are fixed values or tied to the user session, never read from the record.
Unlike the MARC21 records described above, a manuscript or a rare book is catalogued directly in the IFLA LRM model, from the collapsible "Manuscript description"/"Rare book description" sections of the Manifestation form (plus one dedicated field on the Item form): no cascade from a MARC21 record, because the MARC21 standard was not designed to describe support, collation, script, binding or fingerprint at this level of codicological detail. Agents, places and subjects remain linked exactly as in the rest of the catalogue; concepts that are purely codicological, with no dedicated RDA or BIBFRAME element, instead use MYLOD's own local namespace — a deliberate choice: forcing them into a general-purpose vocabulary would make them less precise, not more standard. Two cases deserve a note: a composite manuscript can contain several different texts in the same physical object, each with its own description (incipit, explicit, rubric, colophon, position within the leaves) attached to the single Expression↔Manifestation pair, not to the Manifestation as a whole; when those texts occupy originally independent booklets later bound together, each booklet can have its own codicological unit with its own support, script and date, possibly different from the rest of the volume.
This table has none of the "mandatory/automatic" badges of the MARC21 cascade above: none of these fields is required for a valid record, and none links to an already-existing authority — they are all descriptive data enriching a Manifestation or Item that already exists.
| Field | What it represents | What it generates/links on the LOD side |
|---|---|---|
| ms_supporto | Support material (parchment, paper, mixed, other) | Manifestation attribute; if Parchment or Paper, also linked to the matching Getty AAT concept — no RDA element for support material at this level of detail, MYLOD local namespace. |
| ms_fascicolazione, ms_rigatura, ms_specchio_scrittura (+ mm), ms_righe, ms_numerazione_carte | Quire structure, ruling, written-area dimensions, lines per page, foliation system | Descriptive Manifestation attributes (MYLOD local namespace). |
| ms_scrittura, ms_decorazione, ms_postille | Script type (with guided suggestions), decoration, later-hand marginal additions | Descriptive Manifestation attributes (MYLOD local namespace). |
| ms_legatura, ms_legatura_tipo, ms_legatura_materiale + data | Description, type and material of the binding, with its own dating | Manifestation attributes plus a dedicated Time-span: the binding is often later than the text it encloses, so it has a date independent of the Work's. |
| incipit, explicit, rubrica, colophon, carte_da/carte_a | Description of a single text contained in the Manifestation — a composite manuscript can contain more than one | Dedicated node for each Expression↔Manifestation pair: essential for composite manuscripts, where several different texts share the same physical object, each with its own location. |
| unità codicologica | Originally independent booklets, later bound together into a single codex, each with its own physical description | One node per unit, linked to the Manifestation — the composite structure is described without creating a new Manifestation for every booklet. |
| antico_formula_collazionale, antico_marca_tipografica, antico_impronta, antico_istc_id, antico_edit16_id | Typical elements of a printed rare book description | Manifestation attributes; ISTC and EDIT16 become structured identifiers, like ISBN and ISSN — not plain text. |
| segnature_precedenti, stemma_araldico, fondo, fonte_informazioni, provenienza | Former shelfmarks, ownership marks, collection fund, source of the description, historical chain of ownership | Item attributes plus one event node per ownership transfer (owner, period, acquisition method) — far more queryable than a plain text note. |
For manuscripts, the platform pairs the RDA/LRM catalogue description with a transcription module compliant with the Text Encoding Initiative (TEI P5). A first version of the document — a teiHeader with its msDesc — can be generated automatically from the data already entered in the RDA record, with no extra input required. Alternatively, a transcription produced elsewhere (for instance with a dedicated XML editor) can be imported: the file is strictly validated against the official TEI RelaxNG schema before being accepted, and on error the platform points to the exact line where the XML structure departs from the standard. Each transcription is linked to a specific Expression–Manifestation pair (the individual witness), a requirement for correctly describing a miscellany — a manuscript that gathers several different works.
Every Manifestation can have its own digital copy in PDF format, uploaded directly to the platform. The file can be read from the Manifestation's public record through a built-in viewer, with no option to download or edit it: a choice consistent with its intended use — consultation for study — rather than distribution of the file itself. The platform's general visibility rule still applies here: a digital copy uploaded to a private record in someone's personal space remains visible only to its owner.
A cataloguer working in their personal space is free to reuse an Agent, Place, Subject or Occupation already present in the general catalogue — exactly as they would consult any shared authority file. The entities of the WEMI chain (Work, Expression, Manifestation, Item) follow a different, stricter rule: even when they describe a publication already present in the general catalogue, they always remain distinct and private, and are never merged or overlaid onto existing records. This separation is automatic and requires no special care from the cataloguer: the platform itself guarantees, cascade after cascade, that data produced in the personal space never alters the descriptions held in the general catalogue.