Seg-source and source differences

Looking at some example available for 1.2:
http://docs.oasis-open.org/xliff/v1.2/cs02/Sample_AlmostEverything_1.2_strict.xlf

I've noticed that we don't have much information about how to deal with pre-segmented text units. For example, the document above
contains the following:

<trans-unit id="_1_ski_355">
 <source xml:lang="en">First  sentence.Second sentence.</source>
 <seg-source>
  <mrk mtype="seg" mid="1">First sentence.</mrk>
  <mrk mtype="seg" mid="2">Second  sentence.</mrk>
 </seg-source>
 <alt-trans mid="2" match-quality="75%">
  <source>The second  sentence.</source>
  <target>The translated second  sentence.</target>
 </alt-trans>
</trans-unit>

Here the text content of <source> and the one of <seg-source> is actually different, even after normalizing whitespaces you would
have one space difference.

What the user should do when there is a discrepancy between both content?
Maybe we need more processing descriptions in 2.0?

Cheers,
-yves