There are three different types of letters which appear within the WWP textbase, and each needs to be encoded slightly differently. This entry summarizes our practice in each case. For examples, see the separate entry on each topic.
1. Collections of letters
When a letter appears within a collection which consists entirely of letters, or in an anthology-like collection consisting of letters, essays, and other short works each of which can be encoded as a separate <div>, the WWP uses <div type="letter">. Within <div>, most letters will contain an <opener>, a sequence of body paragraphs, and a <closer>.
Note that <opener> and <closer> have the same internal structure, to accommodate the fact that some letters have the dateline at the end rather than at the beginning. Note as well that <signed> is not only for actual signatures, but for printed names as well.
Note also that the TEI currently does not provide in any satisfactory way for postscripts. The WWP uses the WWP-invented element <ps>.
2. Letters within other works
Because of difficulties with the content model of <div>, there are some places (e.g. within a chapter of a novel) where letters are found but where <div> is not allowed. In such cases, to get around this problem we will nest the <div type="letter"> inside its own <text> and <body> elements. At present this set of nested elements also needs to be nested within a <p> element in order to parse. See the detailed example in the entry on letters within other works.
3. Letters as prefatory material, advertisements, etc.
For letters which are serving another structural purpose (such as prefaces or dedications), we will use <div type="[whatever the thing is. e.g. prefatory]">, and inside use the appropriate letter components: <salute>, <p>, <closer>, etc. without an enclosing <div type="letter">.