More structure for the "name" element in the teiHeader

[christofs]

Our current "name" element for authors and editors has quite a lot of internal structure that is not easy to parse automatically. The author element, in particular, allows strings like this as content:

Lastname, firstname [pseudonym] (birthyear-deathyear)

Could we give more structure to such an entry? I mean, using surname and forename come to mind. Not sure how to solve the remaining information, tbh.

Also, I think it would be important for all of the currently open issues (also: #1 and #2) to keep the schema backwards-compatible, so that the "old" files remain valid against the "new" schema, at least for a certain transitional period.

[lb42]

You mean something like

 <content>
   <alternate>
    <sequence>
     <elementRef key="surname"/>
     <elementRef key="forename" minOccurs="0" maxOccurs="unbounded"/>
     <elementRef key="name" minOccurs="0"  maxOccurs="unbounded"/>
     <elementRef key="date"/>
    </sequence>
    <textNode/>
   </alternate>
  </content>

?

You might also want to require that the pseudonym be distinguished from other names by its @type value, and provide use @from and @to attributes on the <date> That would probably have to be done by schematron, unless you're willing to countenance adding new elements entirely.
What would you do with periphrastic authors l ( "A gentleman", "Two Ladies of Spain") ?
Would you want to distinguish surname and forenames/s inside a pseudonym?

[christofs]

That sounds good!

• Yes, I suppose <name> should have a @type to indicate a pseudonym (or a preferred name).
• Pseudonyms sometimes being just one name (e.g. "Voltaire"), I think the surname/forname distinction is not as important here / might actually cause problems.
• If <name> can be typed, the periphrastic authors could be entered there as well, with an appropriate value on @type. And that's another reason not to distinguish surname/forename there.

[lb42]

Allowing structure within name can be useful in some cases, e.g. you might not know whether "Slingsby" is a forename or a surname. So it seems more consistent to permit it for such cases, Obviously you wouldn't use it for single words (though isn't he sometimes referred to as "M. Voltaire"?)

[lb42]

The TEI Best Practice in Libraries documents suggests (a) we should be using <persName> to wrap <surname> etc. (b) we should also allow for <orgName> as an alternative (c) the usage of multiple <idno> elements to supply multiple authority file references is sanctioned by an example, even though it does seem to contradict the proposed usage for <idno>.