binding: requires a date attribute of some kind @contemporary = 'true'

[mjhawkins]

Is this a rule we want to keep in GS or a MC issue.

[Dr-James-Freeman]

I think @contemporary="true" is problematic in a way that assigning a @notAfter / @notBefore date range is not. Contemporary with what? The date of creation of the manuscript? Or meaning "contemporary with the present era" (aka "modern")? There is an ambiguity there.

A cataloguer might reasonably assume that "contemporary with the creation of the manuscript" is what is intended, but that too risks implying something about medieval book production that we cannot prove, except in very rare instances: i.e. that the binding of manuscripts took place at or shortly after their production had been 'completed' (whatever that means in a culture of bespoke manuscript production).

[mjhawkins]

I can see how the attribute name could be confusing. However, it is the name that TEI chose for the attribute, so we're stuck with it.

The main issue is to decide whether <binding> without at least one of the following attributes notBefore/notAfter/contemporary should trigger an error when you validate. I'm inclined to say that it should not in the core schema. If an institution/project wants to make these items mandatory, they can always do so in their fork of the schema.

All that said. Even if we remove that constraint from the core schema, it doesn't mean that we're washing our hands of these elements. We might decide as a group that it should be recommended to include some sort of 'dating' information about the binding with at least one of these attributes. This can be laid out in the transcription guide - complete with examples. We can also emphasise that contemporary means 'the binding is contemporary with the majority of its contents'.

[ElizabethGow]

I'm generally in favour of providing dating information wherever possible, but think this should be recommended rather than mandatory. When dating information is given, this should be the dating information (notBefore/notAfter/when) AND the contemporary attribute (if a cataloguer can give a date range, they can state if contemporary= "true", "false" or "unknown", and vice versa). This could support a useful browse function, but I don't think should be used to generate displayed text.

[Dr-James-Freeman]

I agree that it shouldn't be used to generate displayed text - the cataloguer can include a prose description of the likely date of the binding, which could be as specific as "16th century" (or more specific) or as general as "medieval" / "post-medieval". I do think that dating of some sort ought to be mandatory, however, since the omission of this data would mean that the manuscript is invisible if someone is filtering manuscripts by date of binding (unless there were also an option for "undated" - for those who like a challenge, perhaps, or want to do their due diligence and leave no stone unturned).

[Dr-James-Freeman]

Also: not an Oxford-specific implementation: we have been using this at the UL.

[mjhawkins]

Leave as it is. Keep as reminder in GS.

The issue of the contents will still need to be discussed.