This is a belated response from a first-time contributor, with a question for the list. David Sewell cited a self-conscious PS at the very top of a letter and asked
> Would anyone argue for encoding it at the end of the document, with a modifier of some sort to indicate place?
Well, maybe. In our on-going edition of the Mary Hamilton Papers (http://www.projects.alc.manchester.ac.uk/image-to-text/), we think logical continuity of content is even more useful than absolute faithfulness to physical layout. Writers sometimes fill every last scrap of white space with continuations of the letter and PSs in various orientations, not to mention leaving space for an address panel which will be on the outside of the folded sheet. Our policy in the XML has been to indicate actual position but to move interruptions of this kind to their logical place, or to the end; a good example is the letter HAM/1/10/1/25.
In the case of the delightful Daniel Webster PS, I'd be tempted to leave it in situ because of its content and because it wouldn't interrupt the text of the letter there. Then it would just be a question of mark-up. (Paul Schaffner's hypothetical but entirely plausible mid-letter PS might be trickier.)
My question is how best to mark material that *is* moved. So far we've been placing an anchor at the original location tied to paired ref tags at the destination location. We're currently revising our mark-up to improve TEI conformance, so now would be a good time to find out whether anchor-ref is the right way to handle this. Advice welcome. Thanks.