The WWP performs only very limited encoding within the content of <mw> elements. The principle involved is that we do not do any content-based or structural encoding eithin <mw>; the only encoding used is to indicate discrepancies (which might be viewed as our error by a user), transcription problems such as illegibility, or basic renditional information.
None of the following elements should be used within <mw> (but note that they are all valid within <mw>):
<placeName>, <persName>, <mcr>, <abbr>, <address>, <date>, <measure>, <name>, <num>, <orgName>, <rs>, <time>, <orig>, <regMe>, <distinct>, <emph>, <foreign>, <gloss>, <mentioned>, <soCalled>, <term>, <title>, <link>, <ptr>, <ref>, <xptr>, <xref>, <mw>, <handShift>
Encoding which we do perform within <mw> is as follows:
1. Where there is an error in the forme work, we encode it using <sic>. For catchwords, we do not include any attributes (e.g. corr=). See 101 for information on the treatment of discrepancies between the catchword and the main text.
2. Where there is highlighting in part of the forme work, we encode it using <hi>. The rendition of an entire <mw> element should be recorded on the rend= attribute of <mw>.
3. If the content of the forme work is obscured, illegible, or deleted, we indicate this as usual using <supplied>, <unclear>, <del>, <gap>.
4. Characters which are usually encoded as delimiters using rend= (e.g. quotation marks, brackets, and so forth) will be encoded as #PCDATA, since they do not function as delimiters for the catchword.