The WWP uses the elements described in section 6.5.3 of the TEI Guidelines to encode handwritten additions and deletions. The only major difference in our usage is that we specify the position of the addition/deletion using the place() keyword on the rend= attribute, rather than the place= attribute specified in TEI. (For more information on how to indicate the position of additions, see the entries on the renditional keyword “place”.) Also, the WWP DTD allows these elements to occur more flexibly than specified in TEI, and in particular allows them to occur within <titleBlock>.
The WWP only encodes handwritten additions which have some linguistic content, by which we mean letters and punctuation. We do not encode non-linguistic additions such as doodles, arrows, or pictures. Our chief reason for omitting hand-drawn pictures even though we do encode printed figures and ornaments is that our primary goal in encoding the text is to capture the printed information. Our encoding of manscript information is cursory at best and we do not have methods at present for accommodating complex handwritten information. Capturing the linguistic information provides the reader with whatever specific commentary accompanies the text.
The elements to use for handwriting are <add>, <addSpan>, <del>, and <gap>.
<add> should be used for any handwritten additions to the text.
<addSpan> should be used to mark the beginning of any longer additions which span several locations.
<del> should be used for any deletions which are still legible. For information on how to handle the legibility of such deletions, see entry on Illegibility and Damage. For deletions which are partly illegible, use <del> with nested <gap> for any completely illegible portions. Use <unclear> nested within <del> to indicate passages which are legible but uncertain.
<gap> should be used for illegible deletions, with a reason=“deleted”.
<gap> should also be used for any part of an addition which is illegible for any other reason, with reason="illegible" (if we believe the text is illegible in the original) or reason="flawed-reproduction" (if we believe the text is illegible due to problems with the photocopy). See 159, 196, and 197 for more detail on using <gap>.
As in all other instances of the resp= attribute, resp= takes the value of the transcriber’s unique name key (for example, “pcaton.xzc”). The cert= attribute is not currently used. The hand= attribute takes as its value the key of the person whose hand we identify; usually this will turn out to be the author. If no one in the office is able to ascertain the hand, the encoder should use the temporary value “unknown.zzx” on the hand attribute, and bring the unknown writing to the attention of one of the staff, who will pass the question on to one of our Board of Scholars. Finally, we do not use the rend= attribute for handwritten words, but use type= instead.
Any keys which are recorded on the hand= attribute for these elements also need to be declared in the TEI header, inside a <handlist> element in the <profileDesc>, immediately after the <langUsage> element. In the following example, two separate people are identified (whose keys appear on the hand= attribute of one or more of the elements described above); one is an unknown person and the other is a fictional person named Mary Fidler.