Dear TC,
The design decision with respect to "related strings" in the Plural, Gender, and Select Module is to represent their "relatedness" via inclusion in a single "unit" element.
<unit id="tu1" pgs:switch="plural:file_count">
<notes>
<note appliesTo="target" ref="tu1_file_count_1"
category="plural_examples">1</note>
<note appliesTo="target" ref="tu1_file_count_few"
category="plural_examples">0, 2~16, 102, 1002</note>
<note appliesTo="target" ref="tu1_file_count_other"
category="plural_examples">20~35, 100, 1000, 10000, 100000</note>
</notes>
<segment id="tu1_file_count_1" pgs:case="1">
<source>You deleted one file.</source>
<target xml:lang="ro">Ați șters un fișier.</target>
</segment>
<segment id="tu1_file_count_few" pgs:case="few">
<source>You deleted <ph id="1" disp="file_count"/> files.</source>
<target xml:lang="ro">Ați șters
<ph id="1" disp="file_count"/> fișiere.</target>
</segment>
<segment id="tu1_file_count_other" pgs:case="other">
<source>You deleted <ph id="1" disp="file_count"/> files.</source>
<target xml:lang="ro">Ați șters
<ph id="1" disp="file_count"/> de fișiere.</target>
</segment>
</unit>
To some, the use of a "group" appears to be an alternative. Rationale:
a. The provenance of multiple segments in a "unit" would be clearer if segmentation is needed
b. The "note" element could be kept closer to the segment that it comments on/annotates
<group id="g1" name="CompositeMessage1">
<unit id="tu1" pgs:switch="plural:file_count">
<notes>
<note appliesTo="target" ref="tu1_file_count_1"
category="plural_examples">1</note>
</notes>
<segment id="tu1_file_count_1" pgs:case="1">
<source>You deleted one file.</source>
<target xml:lang="ro">Ați șters un fișier.</target>
...
</unit>
</group>
Aside: The use of the "group" element was also part of the draft https://docs.oasis-open.org/xliff/v1.2/xliff-profile-po/xliff-profile-po-1.2-cd02.html