FM<-->MIF round-tripping

[Adam Schweitzer]

Is converting between .fm and .mif file formats considered to be safe
(ie. no data lost in translation)?  My understanding is that they should
be - and indeed, I did not notice any glaring differences when I open a
.fm and it's corrosponding .mif in Framemaker.
 
However, when I attempted the following experiment:
1) Open .fm file, save as .mif
2) Open .mif from 1), save as .fm (with different name)
3) Open .fm from 2), save as .mif.
4) Compare .mif files created in 1) and 3) using a text diff tool.
 
I found that there were a number of differences, some of which are
probably safe to ignore (like <XRefLastUpdate> elements), but some are
more concerning (<TabStop> and <PgfNumTabs> elements missing in the
latter file, <PgfPDFStructureLevel> elements having different values..)
 
Goal:
In the process of converting several documents from paper to framemaker
documents, a large number of (rather randomly placed) "discretionary
hyphen" characters were inserted (probably by the OCR software).  I
cannot seem to select these characters for use with the find/replace
tool in Framemaker, so my idea was to convert to .mif, remove all
instances of <Char DiscHyphen>, and then save back as FM.  This appears
to be successful, subject to my concerns above.
 
 
So, I'm wondering.. is it safe to round-trip between .fm and .mif files?
And if not, is there an alternate way of accomplishing my task of
removing these extra characters?
 
Thanks,
Adam Schweitzer
Orenda Aerospace Corporation