I'm currently working on a project that's encoding political cartoons using TEI (using <zone> to create clickable images with pop-up text), and we've run into two big questions relatively early in the project.
1. We'd like to be able to distinguish between an entity being depicted in a cartoon versus being mentioned in the cartoon's text. These are our top contenders for ways to do this, and we'd love feedback about which seems best and/or easiest. (And any
- use <relation> either in the ography files or a separate file that's nothing but a list of relationships.
- put some kind of attribute on <persName> and <orgnName> when they're used in the image encoding. @role? @type?
- represent this relationship in the header of the image xml file
- use <head> and <p> to encode image text in each zone (<head> contains label/ID, <p> contains diagetic text). Then see if we can parse this so that entities tagged in <head> (referencing xml:ids in ographies using <persName>,
<orgName>, <rs>, etc.) are understood to be depicted, and entities tagged in <p> are understood to be mentioned.
2. We want to encode fictional characters and symbols to make a symbol-ography, distinct from the personography of people who actually lived. Is this best done just using <list> or is there some other element we can use or create? Has anyone done this
Digital Projects Associate
Historical Society of Pennsylvania
(215) 732-6200 ext. 205