Open dlschwartz opened 11 months ago
dlschwartz
commented
11 months ago To start the discussion, I propose that we always allow tei:desc as child of tei:relation but that the app only render the tei:desc among the notes when the encoder adds an optional @type="note" attribute.
dlschwartz
commented
11 months ago Here are some examples of different kinds of relationships and how the data about them appears in the TEI.
For relationships with a lot of values and no real need to render data into a prose note, we can include a tei:desc that we wouldn't need to display as a note in HTML. It would display in HTML only in the LOD box. See this example from person/51.
<relation type="place" ref="http://syriaca.org/keyword/has-literary-relation-to-place" source="#bib51-2" active="#person-51"
passive="http://syriaca.org/place/586 http://syriaca.org/place/1468 http://syriaca.org/place/180 http://syriaca.org/place/118
http://syriaca.org/place/174 http://syriaca.org/place/11 http://syriaca.org/place/1469 http://syriaca.org/place/343
http://syriaca.org/place/9 http://syriaca.org/place/10 http://syriaca.org/place/196 http://syriaca.org/place/75
http://syriaca.org/place/1457 http://syriaca.org/place/151 http://syriaca.org/place/663 http://syriaca.org/place/1470
http://syriaca.org/place/1471 http://syriaca.org/place/40 http://syriaca.org/place/122 http://syriaca.org/place/1472
http://syriaca.org/place/215 http://syriaca.org/place/1473 http://syriaca.org/place/1474 http://syriaca.org/place/1475
http://syriaca.org/place/233 http://syriaca.org/place/130 http://syriaca.org/place/65 http://syriaca.org/place/185
http://syriaca.org/place/195 http://syriaca.org/place/1476 http://syriaca.org/place/146 http://syriaca.org/place/211
http://syriaca.org/place/219 http://syriaca.org/place/1477 http://syriaca.org/place/60 http://syriaca.org/place/351
http://syriaca.org/place/89">
<desc xml:lang="en">This author has a literary connection to places.</desc>
</relation>The downside to this is that we lose a meaning rdfs:label for each of these triples associating Severus with a place. The alternative would be to break this out into 37 tei:relation elements with a tei:desc in each that lists Severus by name/URI and the place by name/URI. Either way, we wouldn't want to include this in the list of notes in HTML.
For relationships with a single value we still might not need to render a tei:desc into an HTML note. See this example, also from person/51. BTW, in order to produce a meaningful rdfs:label this tei:desc ought to be rewritten to include <persName ref="http://syriaca.org/person/51>Severus</persName>" instead of leaving him unnamed. Either way, it is probably sufficient to have this data displayed in HTML only in the LOD box.
<relation type="disambiguation" ref="http://syriaca.org/keyword/distinct-entities" mutual="http://syriaca.org/person/51 http://syriaca.org/person/164">
<desc xml:lang="en">Not the same as <persName ref="http://syriaca.org/person/164">Severos</persName>.</desc>
</relation>For cases where we want the note to display in HTML, we add a @type="note" to the tei:desc. Winona would use this to render this tei:desc among the other tei:note elements.
<relation type="disambiguation" ref="http://syriaca.org/keyword/distinct-entities" mutual="http://syriaca.org/person/160 http://syriaca.org/person/444" resp="http://syriaca.org">
<desc xml:lang="en" type="note">
It is unclear why Barsoum asserts that this apparently
West Syrian <persName ref="http://syriaca.org/person/160">Denha</persName>
is the same as an East Syrian <persName ref="http://syriaca.org/person/444">Denha</persName> who wrote a
commentary on the Psalms and several other works. His placement of this
West Syrian Denha in the 9th century does not have any apparent basis
besides this identification, and should also be held loosely.</desc>
</relation>
What use do we have for tei:desc as child of tei:relation?
We aren't currently doing much with most of these. They aren't rendered in the HTML. See person/51 in master: https://github.com/srophe/srophe-app-data/blob/master/data/persons/tei/51.xml#L607. Notice that this record doesn't include a
<note type="disambiguation">. This data only gets rendered in HTML in the LOD box on the right side of the page.Person/160 (in master), however, includes both a
<note type="disambiguation">and a tei:relation[formerly name but now @type='disambiguation']/tei:desc. The note gets rendered in HTML with other notes and the desc does not get rendered even though the tei:relation data does get rendered in the LOD box.We have (rightly, I believe) decided that maintaining the data in two places as in person/160 is not something we want to do. When we have a prose rendering of a relationship like this we should nest it under tei:relation.
This raises an issue about where we want to require tei:relation/tei:desc and where we want to allow it even if we don't require it.
Addressing this issue also requires us to decide what we want to render into HTML. If we don't render everything, how will we designate what we want to render?
<note type="disambiguation">down into the tei:relation/tei:desc: https://github.com/srophe/srophe-app-data/blob/2023.02.21_schemaWork/data/persons/tei/160.xml#L303. I think we would want to render this as a discursive note along with the other notes.