Closed nathangibson closed 6 years ago
@nathangibson I think we need to make some changes to the Zotero to TEI transform to handle the different types of notes. For example in 1.xml I see:
<note type="tag">Subject: Generalia</note>
<note type="tag">!no abstract</note>
<note type="tag">ID: 1</note>
<note type="tag">MS: Berlin, Staatsbibliothek, Violet fragment</note>
<note type="tag">MS: Cambridge, CUL, T-S Ar. 1a.55.</note>
<note type="tag">MS: Istanbul, Süleymaniye, Carullah 3</note>
We could add the subject to the teiHeader//keywords ? The id should probably be stripped as unnecessary? I'm not sure what the best way to handle the MS notes. We could create URIs and add relations but there may be an easier way to deal with them as they are not yet separate entities. If nothing else they could be transformed as note/@type='MS'. Would it make any sense to encode it as a nested bibl, similar to how we handled them in NHSL?
<listRelation>
<relation ref="dcterms:references">
<desc>Berlin, Staatsbibliothek, Violet fragment</desc>
</relation>
</listRelation>
<profileDesc>
<textClass>
<keywords scheme="#lcsh">
<list>
<item></item>
</list>
</keywords>
<classCode scheme="#lc"></classCode>
</textClass>
</profileDesc>
Are you using a controlled vocabulary? If so we can reference it in the scheme attribute.
Currently the zotero2bibl library outputs an incremental idno for the 'local' id and a zotero idno
ex: http://zotero.org/groups/538215/items/EMUXFUUM
I think listRelation method of adding manuscripts is okay, nothing better is springing to mind at any rate.
What do you think of the following relation format for the tags that will go in the subject and manuscript boxes? (It occurs to me that some of the subjects are persons, and some are even other bibliographic items.)
<!--Keywords-->
<relation active="http://biblia-arabica.com/bibl/1" ref="dc:subject">
<desc><term>Historic Printings</term></desc>
</relation>
<!--Persons-->
<relation active="http://biblia-arabica.com/bibl/1" ref="dc:subject">
<desc><persName>Solomon Schechter</persName></desc>
</relation>
<!--Other Bibliographic Items-->
<relation active="http://biblia-arabica.com/bibl/1" ref="dc:subject" passive="http://biblia-arabica.com/bibl/2">
<desc><bibl>Kirsten (1608)</bibl></desc>
</relation>
<!--Manuscripts-->
<relation active="http://biblia-arabica.com/bibl/1" ref="dcterms:references">
<desc><bibl>Berlin, Staatsbibliothek, Violet fragment</bibl></desc>
</relation>
As we assign URIs to these things (probably in a year or two), we could simply add a @passive
to the relation.
@nathangibson I like this approach for keywords. URIs would be great, but we can also facet on the desc/term as long as you are consistent in data entry.
It is not too hard to assign URIs--if they don't have to resolve to anything yet?
No, they are just ids, and more likely to be consistent then free text.
OK, then I think you can write this with the assumption that we will use URIs (as passive values). I will work on minting some--for now there isn't good data for you to test on though.
How would you suggest inputting the data on the Zotero side? What about tags something like: Subject: Historic Printings [http://biblia-arabica.com/keyword/historic-printings] Subject: Solomon Schechter [http://biblia-arabica.com/person/1] Subject: Kirsten (1608) [http://biblia-arabica.com/bibl/1] Subject: Berlin, Staatsbibliothek, Violet fragment [http://biblia-arabica.com/manuscript/1]
So anything with "Subject: " prefix should parse the following text as the relation desc and the URI (in square brackets) as the passive URI?
In case it helps, I've tried this in this example: https://www.zotero.org/groups/538215/biblia_arabica/items/itemKey/4ZD2I7PH
Done.
Right now we are using tags with the format "MS: ..." to indicate which manuscripts are relevant to a particular bibliographic item.
We would like these to appear as subjects on the page, but in a separate box from the subjects.
We will want to be able to browse and search by manuscripts (I will make a separate issue), and later we hope to build a database of manuscripts, but this is a ways off. With this in view, if you think it would be better to create URIs for manuscripts now, let me know and we can do that rather than using the shelfmark string.