Re: [jbovlaste] berbere, berberi

[John Cowan <cowan@mercury.ccil.org>]

Jorge Llambías scripsit:

> Surely that can't have been the reason. How did you prove that
> decompositions with CCVCy- rafsi were unambiguous, and why wouldn't
> the same method of proof work for CCVVCy-? The proof seems just as easy.

Our proof engine was Nora, who reduced the existing morphology to a
set of cases.  Looking at them, we were reasonably sure the CCVVCy
set didn't break the proof, but Nora didn't have time to check it.
Those cases weren't published anywhere I know of, and are either lost
or buried in Bob's paper archive, which amounts to the same thing.
If you want to produce and publish such a proof, I'm sure Lojbanists
everywhere would be happy to have it.

> > Note that a PEG grammar, unlike a YACC grammar, does not prove this,
> > because PEG grammars silently override ambiguities using the rule
> > "first = best".  That isn't good enough for Lojban morphology,
> > so I consider the CLL proposal a dead letter.
> 
> The YACC grammar doesn't handle morphology at all, so I fail to see
> how this has anything to do with the issue.

I was speaking generally.  YACC grammars can provide a proof of
unambiguity in certain circumstances.  PEG grammars cannot: they are
not in that business.

> If the PEG morphology isn't good enough for Lojban, then Lojban doesn't
> have a formal morphology at all, because the PEG is the only one we
> have so far.

The fact that the morphology hasn't changed means that the old proof
is still valid, even though we no longer have access to it.  That's an
unfortunate state of affairs, but it's where we are.

-- 
John Cowan    http://www.ccil.org/~cowan   <cowan@ccil.org>
    "Any legal document draws most of its meaning from context.  A telegram
    that says 'SELL HUNDRED THOUSAND SHARES IBM SHORT' (only 190 bits in
    5-bit Baudot code plus appropriate headers) is as good a legal document
    as any, even sans digital signature." --me

_______________________________________________
jbovlaste mailing list
jbovlaste@lojban.org
http://mail.lojban.org/mailman/listinfo/jbovlaste