As I said in my last email to the list, I'm working on a project in  
which I have a bunch of TEI "manuscript" files, and I'm trying to  
produce a critical edition from them.  I've run into one apparent bug  
and one conceptual problem on which I'd like advice from anyone who  
has done this sort of thing before.

The bug is in the definition of <witDetail/>.  The guidelines say this  
about its attribute 'wit':
"@wit  (witnesses) indicates the sigil or sigla for the witnesses to  
which the detail refers."
...but the RNG schema only allows a single data pointer as the value  
of this attribute.  This is in contrast to <rdg/> and its friends,  
which allow multiple sigla to be listed (e.g. <rdg wit="#A #D  
#Cz">...</rdg>).  Or am I misunderstanding something?  I've modified  
my local copy of the schema, but wonder if this should be fixed in the  
published ones.

Among other more canonical uses, I am using the <witDetail/> tag to  
record punctuation variants.  Rather than making each punctuation mark  
a separate reading, which I have found extremely cumbersome from the  
point of view of text collation, I have been adding a <witDetail/> tag  
on any word that has punctuation in any of the manuscripts.  So a  
typical reading will look something like this (apologies for the  
strange words; this is transliterated Armenian):

      <app xml:id="App481">
          <rdg wit="#Jer #K">
          <rdg wit="#V #Y">
          <witDetail target="#App481" wit="#K" type="punctuation">:</ 
          <witDetail target="#App481" wit=" #V #Y"  
          <rdg wit="#X" type="omission"/>

which gives the following witnesses:
Jer:	zastoucoy
K:	zastoucoy:
V:	astoucoyn,
X:	(om.)
Y:	astoucoyn,

So before I go too much farther with this sort of thing, I am  
wondering: is there a generally accepted way to handle punctuation  
variations in text criticism?  To handle punctuation at the level of  
semantic markup at all?  How have other people done this, if at all?