From MASTER to TEI P5

5. Attribute changes

For reasons elaborated in Driscoll's article, the attributes @type, @status, and @dateAttrib, available for the elements <msDescription> and <msPart> in the MASTER DTD, are removed in the P5 specification. The @type value in MASTER should be transferred to the @form attribute on the new <objectDesc> element, if present. (It is however an open question whether other typologies might be useful in ENRICH). The values for both @status and @dateAttrib proposed in MASTER duplicate information provided elsewhere in the manuscript description and may therefore be ignored.

The following elements from the TEI P5 manuscript description module are typed, that is, they inherit a @type and @subtype attribute from the class att.typed. It would be useful to agree on which of these should have closed sets of values, and what those value lists should be: <accMat>, <altIdentifier>, <custEvent>, <decoNote>, <explicit>, <filiation>, <finalRubric>, <head>, <incipit>, <msName>, <quote>, <region>, <rubric>, <seal>, <settlement>, <stamp>

Similar considerations apply, of course, to many other elements which might be included from the core, transcription, and names and dates modules, notably <name> (for which MASTER already proposes a rather limited value list (other|org|place|person|female)).

The same applies to those elements in MASTER which have a locally defined @type attribute, viz <dimensions>, <ptr>, <head>, <note>, <rubric>, <msName> (formerly <altName>)

Question: for which elements should closed value lists be defined? How should existing data be brought into line with those value lists?

Resolution: The ENRICH project felt that where closed or semi-closed value lists could be easily defined, they should be to encourage consistency, where such lists might prove extremely controversial or had very large number of values they should not be closed. Where we can find possible common values for a (semi-)closed list we should propose it and allow them to disagree if desired.

The @targets attribute on <locus> is renamed as @target in TEI P5. However, it is debatable whether this should be used, given the availability of the new @facs attribute in TEI P5.

Question: should the global @facs attribute be used more widely in ENRICH?

Resolution: The ENRICH project decided, where appropriate information is available in the legacy content, the new TEI facsimile elements and/or attributes should be used where possible.

At TEI P5 there are a number of changes in the globally available attributes. These changes affect all elements. Briefly:
  • @id becomes @xml:id
  • @lang becomes @xml:lang and must specify a valid language identifier (as further defined in P5)
  • @rendition is available as a more precise alternative to @rend.
  • @TEIform is no longer used

There are also, of course, several changes in non-specifically manuscript elements. The attributes of <ptr> and <ref> are different, for example. The content model and attributes of <change> have changed. The ‘mirror tags’ <corr> and <sic> etc. have been replaced by <choice>.

The specialist attributes @technique, @quality, @figurative on <decoNote> are not available in TEI P5. Non-default values for these might be used to develop a typology for the @type attribute. (Manuscriptorium additionally has @illustrative and @size attributes for this element which will need to be handled in the same way).

The @class attribute on <msItem> in MASTER could supply multiple pointers but in TEI P5 may supply only one value (though since this is defined as data.code, the change is probably invisible).

The @langKey attribute on <textLang> is renamed @mainLang. It and @otherLangs now have values which have the same constraints as @xml:lang.

All datable elements are now treated in the same way. In addition to the @notAfter and @notBefore attributes which the MASTER dtd provided, the attributes @when (for an exact point date), @from (for an exact point start date), and @to (for an exact point end date) are now available. These attributes must all supply dates in a valid W3C format (ISO is also an option). An additional attribute @period may be used to indicate a named period as an alternative. The following elements in the manuscript description module are members of this class: <acquisition>, <binding>, <custEvent>, <origDate>, <provenance>, <seal>, <settlement>, <stamp>.

It would be desirable to agree on a common way of normalizing all dates, choosing a subset of these attributes.

Resolution: The ENRICH project decided that where feasible all of the new TEI P5 dating attributes should be used, (i.e. @when, @from and @to in addition to @notBefore and @notAfter and these should be correctly regularised to W3C format.

In MASTER, the elements <acquisition>, <custEvent>, <origDate>, <binding>, <origin> and <provenance> all included two additional attributes (@certainty and @evidence) to indicate the degree of reliability associated with a dating or event. In TEI P5 the equivalent attribute class (att.editLike) provides attributes @cert (equivalent to @certainty), @resp, @evidence and @source. Of the elements listed, only <origDate> and <origin> are members of this class in TEI P5.

Question: Are these indications of uncertainty needed for <acquisition>, <custEvent>, <binding>, or <provenance>? Which of the available attributes should be used?

Resolution: The ENRICH project feels that the new @cert should be used for the old @certainty and other TEI attributes used where possible. The ENRICH project would strongly suggest that the TEI include these attributes on the suggested elements.

At TEI P5, a new module provides for very detailed and consistent annotation of names of places and persons, and of the events and properties associated with them. Some of the attributes used in MASTER have changed as a consequence of this. In particular: on <name>, <country>, <institution>, etc. @key is renamed as @ref, and points to a <person> or <place> element as appropriate. A new @key attribute is used to provide a coded value for such an entity (as an alternative). The @full, @reg and @role attributes are no longer available on these elements. @reg is not available on <origPlace> either.

The measurement elements <dimensions>, <width>, <height>, and <depth> in MASTER have attributes @units and @scope. In TEI P5, @units is renamed @unit, @scope is retained unchanged, and a new attribute @quantity may be used to supply the actual measurement, as an alternative to including it within the content of the element.

It would be desirable to present all measurements uniformly: which method should be adopted?

Resolution: The ENRICH project has decided that where possible all measurements should be converted to millimeters, and where possible this should be stored as a @quantity and @unit. The ENRICH project felt that the current provisions of <dimensions> was limited, and perhaps needed a @shape attribute or similar which would then change the semantics of child elements. (i.e. @shape='circle' means that <width> means 'diameter'). But the ENRICH project believes the TEI should revisit the content model of <dimensions> to record the dimensions of more oddly-shaped objects.

At TEI P5, the schema must use the TEI namespace.

Up: Contents Previous: 4. Elements removed Next: 6. The Manuscriptorium DTD