Done.

It's probably better not to do this on the string level but on some other level,
but I don't really remember the relevant code in the translator. Let me have a
look... something like this:

$ hg diff
diff -r 567207f32fa9 src/translate/fact_groups.py
--- a/src/translate/fact_groups.py	Sat Feb 23 05:59:25 2013 +0100
+++ b/src/translate/fact_groups.py	Fri Apr 19 19:34:49 2013 +0200
@@ -84,7 +84,10 @@
     group_keys = []
     for group in groups:
         group_key = [str(fact) for fact in group]
-        group_key.append("<none of those>")
+        if len(group) == 1:
+            group_key.append(str(group[0].negate()))
+        else:
+            group_key.append("<none of those>")

However, there is some code that relies on the fact names and that will break
given this, function get_predicate_for_fact in
src/search/landmarks/landmark_factory_rpg_sasp.cc. This reliance on fact names
is a major no-no and should be addressed at some point, but at the moment I
think it'd be fine to just patch the code so that it can treat "NegatedAtom"
like "Atom".

Can you make these two changes (translator + search)?

It would be nice to change this more generally to make the
output more readable: if a SAS+ variable has only two values and one of these is
a "none of those", then check if the other is "Atom ..." and call the other
value "NegatedAtom ..." instead.