Redlining/Comparison strategies/COMP problem?

[Bill Troop billtroop@xxxxxxxx]

I wonder if anyone could help me strategize a solution to the
following problem.
1. I have a novel manuscript that I sent to a publisher a year ago.
Having finally accepted the manuscript, the publisher would like a
redlined version showing all the changes I have made since then.
There isn't a huge amount.
2. I write in XyWrite, but I format using the .xtg formatting
language of Quark so that I can easily typeset the MS as I would like
to present it. This is fairly simple, such as  for italic, etc.
etc. Otherwise, it's plain DOS text all the way. [Except of course
for the occasional accented word - - which I fix in Quark, never
having figured out how to get XyWrite to work with Win encoding.]
3. What I would ideally like is something that, like MSW, would take
my two existing documents and (a) mark all insertions in red
(presumably using the Quark code) and (b) mark all deletions in red
and strikethrough.
I can see doing this with Jumbo's COMP command. All I would have to
is find and replace the redlining codes with the Quark codes I want.

4. Alternatively, another solution I can think of would be to:
(a) export the two documents from Quark to MSW (which would remove
all the Quark formatting codes, turning them into their MSW
equivalents), then run and print an MSW redline.
The disadvantage here is that I wouldn't be able to format the
document back in Quark exactly as I would like. But I could survive that.
Any thoughts as to which approach would be better or more foolproof?
Pitfalls to look out for?
I would prefer the COMP approach, except that I am not convinced it
is bulletproof. I have noticed one problem in comparing the two
documents. In one section there is only one change, that the word
'limpid' is inserted. COMP doesn't like this. Instead of putting
limpid as it should, it puts , limpid. The
[998] is highlighted and the word limpid is replaced by a visual
space, except in expanded mode.  continues for several pages,
until the next difference, at which point all is apparently correct.
I didn't notice any problem using XyWrite's built-in compare command
in this passage. It's only COMP that seems to hiccough - - any
particular reason why?

Anyway, it makes me a little nervous about COMP.

What should I do ? ?

Many thanks to all

Bill