[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [bpfk] BPFK work
Jorge Llambías wrote:
On Mon, Oct 11, 2010 at 2:28 PM, Robert LeChevalier <lojbab@lojban.org> wrote:
3. The formalization of the machine grammar has not to my knowledge ever
been fully implemented,
It has for several years now. It hasn't yet been made official, but it
has been formalized, the whole thing, from the morphology up,
including all the tricky magic words.
Where is the parser that handles all the tricky magic words and all the
morphological analysis?
I *said* that the formalization had never been fully implemented.
As far as I am concerned, a formalization which is not implemented has
not been, and cannot be, properly tested to see if the formalization is
"correct".
There are some issues still to decide, but not because they haven't
been formalized, only because we have to make choices as to which
formal version of a rule is preferrable.
If we haven't decided, then there is no complete formalization, much
less an implementation of the formalization.
and the fact that we are having this discussion
indicates that the concept of "text" was never formally defined in terms
necessary to account for multiple speakers.
That's a pragmatic matter, external to the formal grammar, right.
That depends on whether we formalize pragmatics.
By the formal language, no fa'o: no end of text.
That's not quite right.
In PEG, FAhO is optional, otherwise the end of input is enough to
signal the end of text.
How does the computer know "end of input"? "fa'o" is the equivalent of
whatever signal the computer uses to detect that state.
In BNF the rule for FAhO is only given informally, it is not a part of
the formalization:
"FAhO is a universal terminator and signals the end of parsable input."
Then the BNF is not a complete formal grammar.
By the formal grammar, if there is no fa'o, there is no new text. Indeed, by
the formal grammar, there is no concept of more than one text and after a
fa'o, everything else until the end of time is non-Lojban. %^)
That's absurd. By the formal grammar, each input that parses correctly
is a valid Lojban text. There's no reason to think that only one valid
Lojban text exists in all the universe.
There is no reason NOT to think so. In a formalism, only that which has
been formalized exists.
Why? Why wouldn't the computer just parse each speaker's text on its
own?
The computer doesn't know what a "speaker" is.
Computers nowadays are pretty good at voice recognition.
If the formal grammar specifies the decisions of a particular voice
recognition program as constituting the definitions of multiple texts,
then that will be relevant.
If there is no specified formal rule, then it doesn't exist as part of
the formal system.
It knows what a text-stream
is, and that such a text-stream ends with a fa'o and only with a fa'o.
You can program a computer to recognize different speakers and feed
the contribution from each speaker to the parser as a separate input.
That of course has nothing to do with the formal grammar.
That programming would in fact become part of the formal grammar.
Absurd.
No. Just the limitations of formal systems.
lojbab
--
You received this message because you are subscribed to the Google Groups "BPFK" group.
To post to this group, send email to bpfk-list@googlegroups.com.
To unsubscribe from this group, send email to bpfk-list+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/bpfk-list?hl=en.