Dear Hayim,
intrigued by your post I did reread relevant section and I wonder what is it that you are interested in: the metamarks or the transpositions (or both)? 

Looking at transposition examples in the Guidelines it looks to me that positioning of the <metamark> element could be easily as a child or a sibling of the relevant passage (but definitely not limited to these positions, given the broad range of proofing metamarks in use). Nevertheless for reconstituting the intended order of the text the only relevant part are listTranspose/transpose entries, not the metamarks. Key is to have something to point <transpose> to, so as you mentioned to use <w> if fitting or dedicated <seg> as identifiable wrapper element for the chunk being transposed.



On Wed, 26 Dec 2018 at 22:02, Hayim Lapin <[log in to unmask]> wrote:
Happy holidays to all.
I am wondering how those of you who work with MSS/genetic editions handle transpositions. 
<metamark>: assuming the unit being transposed is a <w> would you put the metamark inside the w or outside (as the example from the docs do it)? 
My instinct is that it is not significantly different from any other scribal intervention (deletion or addition of characters), and should be included in the markup of the word. (It may be important here that for alignment and other processing we segment the text into w elements each with an xml:id.

Wrapping the transposition: This is relates more to processing than encoding, but the expression is in the tagging. My first impulse is to wrap all the affected transposed material in a parent tag: in-word characters in a w element, word or multi-word-chunk transpositions with a single seg. The wrapping element has the type "transp" or something like that. 
This seems useful when handling the use case where one wants to be able to show the text at specific points (in this case before and after transposition.) As with add|del|damage the xslt|xq|etc can receive specific instructions when it encounters such a seg (or w): use the appropriate "transpose" to swap elements. 
The alternative would be that when we handle a particular text container (ab, in my case) we check if it has transposed text and swap the the text first before doing other processing.

 Sorry for this long and possibly confusing email,


Hayim Lapin
Professor of History
Robert H. Smith Professor of Jewish Studies
University of Maryland
Jewish Studies: 4141 Susquehanna Hall, College Park, MD 20742 | 301 405 4975
History: 2115 Francis Scott Key Hall, College Park, MD 20742 | 301 405 4296