The richness and size of the header reflect the diversity of uses to which electronic texts conforming to these Guidelines will be put. As described in section , the TEI header allows for the provision of a very large amount of information concerning the text itself, its source, encodings, and revisions as well as detailed descriptive information that can be used by researchers in analysing the text. The richness of the header will depend on the nature and intended use of the text. At one extreme, an encoder may expect that the header will only provide bibliographic information about the text adequate to local needs. At the other, wishing to ensure that their texts can be used for the widest range of applications, encoders will want to document as explicitly as possible both bibliographic and descriptive information in such a way that no prior or ancillary knowledge about the text is needed in order to process it. The header, in the latter case, will be very full, approximating the kind of documentation often supplied in the form of a manual. Most texts will lie somewhere between these extremes; textual corpora in particular will tend toward the latter extreme.

For a successful and complete mapping from the TEI header to other standards and formats, the TEI header should be as complete as possible. When deciding which information to include in a TEI header that will be mapped to another metadata standard and the format or structure of that header information, the following should be kept in mind: The header should provide full bibliographic information on the encoded text, its source, where the text can be located, and any restrictions governing its use. The header should contain useful information about the encoding of the text itself. In this regard, it is highly recommended that the encoding description be as complete as possible. The Guidelines do not require that the encoding description be included in the header (since some simple transcriptions of small items may not require it), but in practice a header without an encoding description would prove difficult mor impossible to map satisfactorily to most other metadata formats. The header should be amenable to automatic processing, particularly for loading into databases and for the creation of publications, indexes, and finding aids, without undue editorial intervention. For this reason, two recommendations are made regarding the format or structure of the header: first, where there is a choice between a prose content model and one that contains a formal series of specialized elements, wherever possible and appropriate the specialized elements should be preferred to unstructured prose. For instance, the source description can contain either a free-prose citation (tagged bibl or even p) or a biblStruct element, which provides a more rigorous structure for the bibliographic information (see examples in section ). The more structured biblStruct element is more suitable for automatic processing, and is therefore recommended over the less structured alternatives whenever the header is to be mapped to another standard. Second, with respect to corpora, information about each of the texts within a corpus should be included in the overall corpus-level teiHeader. That is, source information, editorial practices, encoding descriptions, and the like should be included in the relevant sections of the corpus teiHeader, with pointers to them from the headers of the individual texts included in the corpus. There are three reasons for this recommendation: first, the corpus-level header will contain the full array of bibliographic and documentary information for each of the texts in a corpus, and thus be of great benefit to remote users, who may have access only to metadata from the corpus-level header; second, such a layout is easier for the coder to maintain than searching for information throughout a text; and third, generally speaking, this practice results in greater overall consistency, especially with respect to bibliographic citations.

The distribution and retrieval of TEI header metadata enable resource discovery by other means. The mappings to MARC and Dublin Core discussed in the remainder of this section provide two examples of how TEI header metadata may be re-used and re-purposed.

This section offers some guidance to both cataloguers and bibliographic analysts who want to load metadata from TEI headers into a MARC-based retrieval system. Because there are variations in cataloguing practice across local sites, among bibliographic utilities (such as OCLC and RLIN), and differences in MARC usage in different countries, only tentative advice is possible. Note that the following examples are based on USMARC, not UNIMARC.For more information on UNIMARC, see Brian P. Holt, (London, U.K.: IFLA Universal Bibliographic Control and International MARC Programme, British Library, 1987). For USMARC, see Walt Crawford, (Boston: G.K. Hall, 1989), (Washington, D.C.: Library of Congress, 1987), and Deborah J. Byrne, (Englewood, Colo.: Libraries Unlimited, Inc., 1991). UNIMARC offers cataloguers in different countries the opportunity to combine different national practices in a single MARC format, and is the preferred variety of MARC records for distribution across national boundaries. The implementation of UNIMARC, however, will be affected by local practice and by guidelines offered by the bibliographic utilities. Though UNIMARC is a stable format, the guidelines for its implementation are not sufficiently known or stabilized to be included in this chapter.

There are some major differences between the MARC record and the TEI header that will cause problems for librarians trying to map from the TEI header to the MARC record. The most important difference between the MARC record and the TEI header is the function of each. The primary function of the MARC record when it was first designed in the mid-1960s was to allow for the electronic distribution of cataloguing records in support of card production. See Henriette Avram, (Washington D.C.: Library of Congress, 1968), p. 3. For discussion of the relationship between the MARC record and the catalogue card, see Michael Gorman, , in Richard Smiraglia, ed., (Chicago: American Library Association, 1992). The catalogue card is a unitary record for a physical object containing complex bibliographic data of varying sorts. The catalogue card points to the physical object. The TEI header provides full bibliographic information (as would a card), as well as documentary non-bibliographic information that supports the analysis, either by humans or machines, of the electronic text documented by header. Most of this analytical information, which is found in the profile description, encoding description, and revision history, has little direct provision for it in the MARC record, and if retained must be recorded as unstructured notes (55XX) fields. Notes fields usually do not have the structure to support machine retrieval and analysis, while properly formatted profile, encoding, and revision descriptions lend themselves to retrieval, can support machine processing (including analysis), and point directly to the electronic text attached to the header. Moreover, the electronic text points back to the relevant elements in the header.

Though this section offers some advice on where the profile, encoding, and revision descriptions might go in a MARC record, for practical reasons a repository might want create a codebook from these divisions of the header, and create a MARC record from the file description only. The MARC record should contain a reference to the codebook.

Subfields (or delimiters) are conventionally indicated by the dollar sign.

<!-- TEI header / Dublin Core discussion to go here. -->