[lojban] Re: [lojban-beginners] A challenge for computer science/programming geeks: The LLG wants to give you $500!

["Stephen Pollei" <stephen.pollei@gmail.com>]

>
> That's not even remotely the same language.  For one thing, there's
>  no way to get to A, so it's actually:
>
>  S => a+ B
>  B => b B? c
>
>  which is just a^mb^nc^n, which is a completely different language.

the weakened cfg clearly parses a superset of the example peg, that is
true and I'm a 100% possitive that if you took you and xorxes peg
things and reduced it to a cfg that it would also parse a superset of
what your peg parses and do it ambiguiously to boot.

>  You may already understand this, and that's fine, I just want there
>  to be no confusion: a language that is as close to Lojban as that is
>  to a^nb^nc^n is of no use to us, and we won't pay out for it.

Yes I never argued that the resultant cfg would be any good merely
that it would parse all valid lojban that your peg already parses.
Also I'd claim that the cfg would produce set of answers that would
include your peg answer for any given valid lojban text which your peg
parses.

Also yes the resulting cfg would parse some text which is invalid
lojban, especially at the word morphology level I see that it has some
validation checks for things like invalid consonant clusters which the
cfg version of it would lose. At the grammar level though I'm unsure
if the extra invalid texts it would parse would: a) exist b) be
noticeable . The one thing I'm certain of is that it would be ambigous
not even handling {boi} correctly.
The {boi} example is really what I thought was more interesting then
the thought that you could weaken any peg.


To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.