Re: Is this ambiguous?

> Here's a grammar that tries to defeat an algorithm based only on
> vertical loops (derivation of N-from-to from N-from-to) by making
> horizontal loops.

Thank you. I had got as far as thinking to myself “it must be possible
to make loops that aren’t just choices off single nodes” but I hadn’t
got as far as trying to construct one.

Looking at the graph[1], it’s clearly going to require a lot more
analysis to work out whether or not it would produce different vxml
output.

It would.

$ coffeepot -g:examples/horiz.ixml -p:all x
<ixml parses="9" totalParses="9">
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <A/>
      <B/>
      <C/>
      <D>x</D>
   </S>
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <A/>
      <B/>
      <C/>
      <D>x</D>
   </S>
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <D>x</D>
      <A/>
      <B/>
      <C/>
   </S>
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <A/>
      <B/>
      <C/>
      <D>x</D>
      <A/>
      <B/>
      <C/>
   </S>
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <A/>
      <B/>
      <C/>
      <D>x</D>
      <A/>
      <B/>
      <C/>
   </S>
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <D>x</D>
      <A/>
      <B/>
      <C/>
   </S>
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <A/>
      <B/>
      <C/>
      <D>x</D>
   </S>
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <A/>
      <B/>
      <C/>
      <D>x</D>
   </S>
   <S xmlns:ixml="http://invisiblexml.org/NS" ixml:state="ambiguous">
      <D>x</D>
   </S>
</ixml>

That’t sufficient for me to put “improving ambiguity detection” on the
back burner indefinitely.

                                        Be seeing you,
                                          norm

--
Norm Tovey-Walsh
Saxonica