one approach to implementing \power macro

lib/LaTeXML/MathParser.pm

@@ -1529,7 +1529,8 @@ sub NewScript {
     $l++; $bumped = 1 }
   elsif (my $innerl = p_getAttribute($rbase, '_bumplevel')) {
     $l = $innerl; }
-  my $app = Apply(New(undef, undef, role => $y . 'SCRIPTOP', scriptpos => "$x$l"),
+  my $meaning = p_getAttribute($rscript, 'meaning');
+  my $app = Apply(New(undef, undef, role => $y . 'SCRIPTOP', scriptpos => "$x$l", ($meaning ? (meaning => $meaning) : ())),
     $base, Arg($script, 0));
   # Record whether this script was a floating one
   $$app[1]{_wasfloat}  = 1  if $mode eq 'FLOAT';

lib/LaTeXML/Package/a11ymark.sty.ltxml

@@ -28,5 +28,21 @@ DefMath('\deriv[]{}{}',
 
 DefMath('\integral{}{}', '\int #1 \diffd #2', meaning => 'integral');
 
+DefConstructor('\meaning{}{}', '#2', afterConstruct => sub {
+    my ($doc, $whatsit) = @_;
+    my $apply = $doc->getNode->lastChild;
+    if ($apply->localName eq 'XMApp') {    # should be an apply, if by design
+      $apply->setAttribute('meaning', ToString($whatsit->getArg(1))); } },
+  bounded => 1, requireMath => 1);
+
+DefMacro('\power{}{}', '\meaning{power}{#1^#2}');
+
+# These still need work, since currently we get
+# annotations "transpose(#1,#2)" and "frobulator(#1,#2)"
+# instead of the desired "transpose(#1)" and "frobulator"
+# I guess this needs a different type of XMath structural pattern...
+DefMacro('\transpose{}', '\meaning{transpose}{#1^T}');
+DefMacro('\frobulator',  '\meaning{frobulator}{x\'}');
+
 ########
 1;